/3. MONOGRAPH OF THE VOLES & LEMMINGS MONOGRAPH OF THE VOLES & LEMMINGS (MICROTIN.E) LIVING AND EXTINCT BY V MARTIN A. C. HINTON LONDON : PRINTED BY ORDER OF THE TRUSTEES OF THE BRITISH MUSEUM SOLD AT THE BRITISH MUSEUM (NATURAL HISTORY), CROMWELL ROAD, S.W. 7 AND BY B. QUARITCH, LTD.; DULAU & CO., LTD.; THE OXFORD UNIVERSITY PRESS; AND WHELDON & WESLEY, LTD., LONDON; ALSO BY OLIVER & BOYD, EDINBURGH 1926 {AU rights reserved) Issued 26th June, 192fi] Frintep in Great Britun by Richard Clay & Sons. Limited, bungay, suffolk. PREFACE The Voles and Lemmings form a group that presents many difficulties to the systematist ; they are also of special interest in relation to the dental evolution of the Simplicidentate Rodents. Mr. Hinton's monograph is the result of an intensive study that began more than twenty-five years ago; his work on the fossil species has led him to certain conclusions as to the sequence of Pliocene and Pleistocene deposits and the date of the Glacial Period, which were put forward at the 1925 meeting of the British Association, and will be published in the next number of the Proceedings of the Yorkshire Geological Society. This first volume includes a general account of the Microtinss and a systematic revision of 14 of the 31 genera; the second, the preparation of which is well advanced, will conclude the work. Attention may be called to Mr. Hinton's views as to the impor- tance of habits and environment in specific differentiation ; when ideas of this kind are derived from a profound and intimate study of a group they should be worthy of serious consideration. The author wishes to acknowledge his indebtedness to Mr. A. C. Savin of Cromer for the loan, without restriction of time, of a very large collection of Microtine remains from the Cromerian Beds; also to Mr. W. J. Lewis Abbott, Miss D. M. A. Bate, Dr. H. P. Blackmore, Dr. Frank Corner, Mr. J. Wilfrid Jackson, Mr. A. S. Kennard, Dr. H. C. Male, the late Dr. C. I. Forsyth Major, Dr. L. S. Palmer, Prof. S. H. Reynolds, Dr. R. F. Scharff, Dr. 0. Wettstein, and Mr. Gilbert White for gifts or loans of fossil material. Thanks are also due to the Geological Survey of England and Wales, the Irish National Museum, the Royal Scottish Museum, the Museums of Cambridge University, Bristol, Norwich, Ipswich, Taunton, and Wells, the Bristol University Spelseological Society and the Vienna Museum, for the loan of material. The author wishes to thank many friends for assistance of various kinds, including Mr. Oldfield Thomas, Mr. R. Oldfield, Mr. B. B. Woodward and Mr. C. D. Sherborn, but particularly VI PREFACE the Hon. Ivor Montagu, who made a special examination of the great collections in Leningrad, and collected important material in the Caucasus. The drawings of skulls have been made by Mr. Terzi; those of the teeth are by the author. The index has been prepared by Miss N. S. Kaye. C. Tate Regan (Keeper of Zoology). British Museum {Natural History), June, 1926. CONTENTS PAGE I. Structure and Classification : («) INTRODUCTORY ...... 1 {h) PREVIOUS WORK AND PRESENT TREATMENT (c) GENERAL CHARACTERS OF THE SUBFAMILY MICROTIN^ ...... 5 {d) THE EVOLUTION OF THE MICROTIN.E AND THE INTERRELATIONSHIPS OF THE GENERA . 26 (e) KEY TO THE GENERA ..... 88 (/) SPECIAL NOTES : 1. THE EVOLUTION OF THE INCISOR TEETH 99 2. THE EVOLUTION OF THE CHEEK-TEETH . 102 3. THE DENTAL FORMULA OF THE MURID^ 124 4. THE RANGE IN TIME OF BRITISH FOSSIL MICROTIN^ ..... 125 II. Sy-stematic Revision of the Microtin.e . . 137 SYSTEMATIC INDEX ORDER RODENTIA. FAMILY MURID^. Subfamily Microtince. PAGE 1. Dicrostonyx, Gloger 137 1. torquatus, Pallas 148 la. t. torquatus, Pallas 148 16. t. ungulatus, von Baer 149 2. chionopses, Allen 150 3. rubricatus, Richardson 151 3a. r. rubricatus, Richardson 151 36. r. richardsoni, Merriam 153 3c. r. unalascensis, Merriam . . ... . . 154 4. exsul, G. M. Allen 155 5. groenlandicus, Traill 155 6. hudsonius, Pallas 157 7. fgulielmi, Sanford 160 8. fhenseli, Hinton 163 2. Synaptomys, Baird . 165 1. cooperi, Baird 169 2. fatuus. Bangs 169 3. helaletes, Merriam 170 3a. h. helaletes, Merriam 170 36. h. gossii, Coues 171 4. borealis, Richardson 172 5. dalH, Merriam 173 6. wrangeli, Merriam 174 7. andersoni, Allen 174 8. chapmani, Allen 175 X SYSTEMATIC INDEX PAGE 2. Synaptomys, Baird (continued) 9. truei, Merriam 175 10. innuitus, True 176 10a. i. innuitus, True 176 106. i. medioximus, Bangs 177 11. sphagnicola, Preble 177 3. Myopus, Miller 178 1. schisticolor, Lilljeborg 181 2. morulus, HoUister 182 3. saianicus, Hinton 182 4. middendorffi, Vinogradov 183 5. thayeri, G. M. Allen 184 6. brandti, Tscherski 185 4. Lemmus, Link 186 1. lemmus, Linnaeus 193 2. obensis. Brants 201 2a. o. obensis, Brants 201 26. 0. novosibiricus, Vinogradov 202 2c. o. chrysogaster, Allen 203 3. amurensis, Vinogradov 204 4. paulus, G. M. Allen 205 5. nigripes, True . 206 6. alascensis, Merriam 207 7. yukonensis, Merriam 207 8. minusculus, Osgood 208 9. trimucronatus, Richardson 208 10. helvolus, Richardson 209 5. Evotomys, Coues 210 1. fharrisoni, Hinton 216 2. glareolus, Schreber 217 2a. g. glareolus, Schreber 218 26. g. britannicus. Miller 219 2c. g. reinwaldti, Hinton 220 2d. g. suecicus. Miller 221 2e. g. istericus, Miller 221 SYSTEMATIC INDEX XI PAGE Evotomys, Coues [continued) 2. glareolus, Schreber (continued) 2/. g. helveticus, Miller 222 2g. g. sobrus, Montagu 223 2h. g. saianicus, Thomas 223 3. centralis, Miller l 224 4. fkennardi, Hinton 225 5. nageri, Schinz 226 5a. n. nageri, Schinz 226 56. n. italicus, Dal Piaz 227 5c. n. vesanus, Hinton 228 5d. n. hallucalis, Thomas 228 5e. n. vasconise, Miller 229 5/. n. norvegicus. Miller 230 6. gorka, Montagu 231 7. skomerensis, Barrett-Hamilton 231 8. csesarius. Miller 233 9. alstoni, Barrett-Hamilton and Hinton .... 234 10. erica, Barrett-Hamilton and Hinton .... 235 11. ponticus, Thomas 235 12. frater, Thomas 236 13. baikalensis, Ognev 237 14. laticeps, Ognev 237 15. parvidens, Ognev 238 16. otus, Turov 238 17. rutilus, Pallas 238 17o. r. rutilus, Pallas 239 176. r. russatus, Radde 240 18. wosnessenskii, Poljakov 240 19. jochelsoni, Allen 241 20. amurensis, Schrenck 242 21. mikado, Thomas 242 22. rufocanus, Sundevall 243 22«. r. rufocanus, Sundevall 248 226. r. shanseius, Thomas 251 22c. r. regulus, Thomas 256 22d. r. smithii, Thomas 257 XU SYSTEMATIC INDEX P&QE 5. Evotomys, Coues (continued) 23. caurinus, Bailey 262 24. phseus, Swarth 262 25. wrangeli, Bailey 263 26. dawsoni, Merriam 264 26a. d. dawsoni, Merriam 265 266. d. insularis, Heller 266 27. orca, Merriam 266 28. gapperi, Vigors 266 28a. g. gapperi, Vigors 267 28&. g. ochraceus, Miller 268 28c. g. rlioadsii, Stone 268 28d. g. loringi, Bailey 269 28e. g. athabascse, Preble 270 28/. g. galei, Merriam 270 28^. g. saturatus, Rhoads 271 29. brevicaudus, Merriam 271 30. carolinensis, Merriam 272 31. limitis, Bailey 273 32. ungava, Bailey 273 33. idahoensis, Merriam 274 34. mazama, Merriam 274 35. obscurus, Merriam 275 36. calif ornicus, Merriam 275 37. occidentalis, Merriam 276 38. nivarius, Bailey 277 39. proteus, Bangs 277 6. Aschizomys, Miller 278 1. lemminus, Miller 279 7. Eothenomys, Miller 280 1. melanogaster, Milne-Edwardes 285 la. m. melanogaster, Milne-Edwardes . . . 285 16. m. cachinus, Thomas 286 Ic. m. eleusis, Thomas 286 Id. m. aurora, Allen 287 SYSTEMATIC INDEX xiil PAGE 7. Eothenomys, Miller (continued) 1. melanogaster, Milne-Edwardes (conlinued) le. m. miletus, Thomas 287 1/. m. confinii, Hinton 288 Ig. m. colurnus, Thomas 288 Ih. m. mucronatus, Allen 289 li. m. libonotus, Hinton 289 2. fidelis, Hinton 290 3. proditor, Hinton 291 4. olitor, Thomas 292 8. Anteliomys, Miller 292 1. chinensis, Thomas 296 la. c. chinensis, Thomas 297 16. c. tarquinius, Thomas 297 2. wardi, Thomas 298 3. custos, Thomas 298 3a. c. custos, Thomas 299 36. c. rubelius, G. M. Allen 299 9. Alticola, Blanford 300 1. argurus, Thomas 308 2. blanfordi, Scully 308 2a. b. blanfordi, Scully 309 2b. h. lahulius, Hinton 309 3. roylei. Gray 310 3a. r. roylei, Gray 312 36. r. cautus, Hinton . . • 313 4. albicauda. True 314 5. montosa. True 315 6. glacialis, Miller 316 7. phasma. Miller 317 8. worthingtoni, Miller 318 8a. w. worthingtoni. Miller 319 86. w. subluteus, Thomas 319 8c. w. semicanus, G. M. Allen 320 9. stoliczkanus, Blanford 320 XIY SYSTEMATIC INDEX PAGE 9. Alticola, Blanford {continued) 10. stracheyi, Thomas 321 11. lama, Barrett-Hamilton 324 12. acrophilus, Miller 324 13. strelzowi, Kascenko 326 14. alliarius, Pallas 329 10. Hyperacrius, Miller 330 1. fertilis, True 334 la. f. fertilis, True 334 16. f. brachelix, Miller 335 2. aitchisoni, Miller 336 3. wynnei, Blanford 336 11. Dolomys, Nehring 339 1. fmilleri, Nehring 341 2. tepiscopalis, Mehely . 342 3. bogdanovi, Martino 343 12. fApistomys, Mehely . . 348 1. fcoronensis, Mehely 349 13. IMimomys, Forsyth Major 350 1. tpliocsenicus, Forsyth Major 357 2. treidi, Hinton 363 3. fsavini, Hinton . . . 365 4. fintermedius, Newton 368 5. tpusillus, Mehely 374 6. fnewtoni, Forsyth Major 375 7. fmajori, Hinton 378 8. fcantianus, Hinton 383 14. Arvicola, Lacepede 385 1. fbactonensis, Hinton 386 2. tgreenii, Hinton 389 3. tprseceptor, Hinton 391 4. mosbachensis, Schmidtgen 394 SYSTEMATIC INDEX XV PAGE 14. Arvicola, Lacepede (continued) 5. amphibius, Linnaeus 395 5a. a. amphibius, Linnaeus 400 5b. a. reta, Miller 401 6. sapidus, Miller 401 6a. s. sapidus, Miller 402 66. s. tenebricus, Miller 40^ 7. terrestris, Linnaeus 403 7a. t. terrestris, Linnaeus 404 76. t. italicus, Savi 404 7c. t. musignani, de Selys-Longchamps . . . 405 Id. t. illyricus, Barrett-Hamilton. . . . 406 7e. t. rufescens, Satunin 406 If. t. meridionalis, Oguev 407 7g. t. persicus, de Filippi 408 7h. t. scythicus, Thomas 409 8. scherman, Shaw 410 8a. s. scherman, Shaw 412 86. s. exitus, Miller 413 8c. 8. monticola, de Selys-Longchamps . . . 414 9. fabbotti, Hinton 414 10. fautiquus, Pomel 417 LIST OF PLATES PLATE I. Palates op Microtus. II. Ondatra zibethica Linn^us. New-born skull. III. Arvicola ampeibivs Linn^us. Skull and limb- skeleton OF an old individual. IV. Arvicola terrestris Linn^us. Skull and limb- skeleton OF AN OLD individual. V. DiCROSTONYX OULIELMI SanFORD. Co-TYPES. VI. Skulls of Lemmus. VII-XII. Skulls of Evotomys. XIII. MiMOMYS. Lateral views of mj. XIV. MiMOMYS INTERMEDIUS NewTON. PaLATE AND mandible. XV. MiMOMYS MAJORI HiNTON. OuTER VIEWS OF »?i. MONOGRAPH OF THE MICROTINJi I. STRUCTURE AND CLASSIFICATION a. Introduction. The Voles and Lemmings are SimpHcidentate Rodents belonging to the great family Muridae, of which they constitute a subfamily, the Microtinaj. They are very widely distributed, ranging throughout the Holarctic region from the limits of terrestrial mammalian life in the high north southwards to the Alpine meadows of the mountains of Guatemala (just within the Neotropical region) and to those of Yunnan and Burma (in the Oriental region). They are not, however, known to live in Northern Africa, although a species of the fossorial genus Ellobius has left its remains in the Pleistocene deposits of Tunis and Algeria, and members of less highly specialized genera inhabit or have inhabited some of the islands of the Mediterranean. The vertical range of the subfamily is also wide, extending from the sea-beach up to the limits of mammalian life, at altitudes of about 20,000 feet in the Himalayas. The oldest forms yet discovered are found in the later Pliocene deposits in Europe, and are already fully developed voles. Not improbably the Microtinse originated in early Tertiary times in Northern and Central Asia, and it is to future palaeontological work in that continent and perhaps in the high north that we must look for information about the forms which once connected the Microtinae with the other subfamilies of Muridse. Having regard to the sedentary habits of these rodents and to their wide distribution, it is not surprising to find that the sub- family is a large one, comprising thirty-one genera and several hundred species. Within rather narrow limits the species show considerable plasticity, developing numerous more or less well- marked geographical or purely local races to meet the exigencies of the environment. They have therefore attracted a great deal of attention from systematists, and a very large number of generic, subgeneric, specific and subspeciflc names have been bestowed upon them in an extensive and widely scattered literature. In consequence the group has become one of the most difficult to deal with from a systematic point of view. Quite apart from this, however, many of the most important and most V I,. B 2 MICEOTIN^ widely ranging species are extraordinarily perplexing animals. Althougli they attain sexual maturity and an adult appearance at a very early age, their growth continues for an unusually long period, during which colour, form of the skull and structure of the teeth may change to such an extent that in old age they may present characters widely different from those shown in the earlier " adult " stages of growth. To work out the life-histories of such mammals thoroughly requires a large amount of material and more time than is at the disposal of a systematist busied with the task of determining collections as they arrive in the Museum. One general result of my work has been to reduce the number of genera and species considerably. The Microtinae are of great general interest and merit careful attention. Their structure is highly modified in relation to subsistence upon a diet despised by most other Muridse, and to their burrowing habits, but it retains certain primitive features which seem to throw light upon the problems connected with the ancestry of the Simplicidentate Rodents and with the evolu- tion of their molars. The close adaptation of form and structure to the needs of special functions shown by the skulls, teeth, and other organs of the Microtinse deserve close study; in one genus {Mimomys) at least, where we can trace a single phylum forward in time, we find a good example of recapitulation. Owing to the short range in time of the individual species the group is of great importance to the geologist who has to correlate scattered deposits like those of the Pleistocene period in Britain. The Microtinse become of considerable economic importance in certain circumstances, as in " mouse years," when, favoured by lenient weather and often by unwitting human intervention, voles and lemmings swarm and inflict great losses upon agrarian industries. In such years, too, riparian species occasionally do serious damage to the banks of streams and canals by their bur- rowing operations. When present in reasonable numbers these rodents, like most other creatures, play their part in maintaining the general balance of Nature, and in uncultivated lands their fossorial activities endow many tracts, which would otherwise be completely barren, with comparative fertility. Some species are apparently concerned in the dissemination of disease, as, for example, the Japanese Microtus tnontebelloi, which is believed to be one of the natural reservoirs of Infectious Jaundice and Japanese River Fever. The North American genus Ondatra is of much value as a fur-bearer. During the open season the flesh of this animal is extensively used as food in the large towns of eastern North America. Much remains to be done before a really satisfactory account of the subfamily can be written. The present work is intended to summarize existing knowledge and to direct attention to some points of more general interest that require further investigation. PREVIOUS WORK b. Previous Work and Present Treatment. Modern knowledge of the Microtinse is based upon Miller's " Genera and Subgenera of Voles and Lemmings," published in 1896.^ In that fine work Miller gave an admirable review of the earlier classifications and established a natural classification which swept away the more or less artificial systems of his predecessors (de Selys-Longchamps 1836-1862, Blasius 1857, Baird 1857, Fatio 1867, Coues 1874, Blanford 1881, Lataste 1887). Subsequent research has tended only to extend and to correct in detail the results at which Miller arrived. Miller confined his attention chiefly to the living members of the subfamily, of which he recognized seven genera, namely, Synaptomys, Lemmus, Dicrosto)>yx, Phenacomys, Evotomys, Microtus and Fiber [= Ondatra]. Synaptomys included two subgenera, Synaptomys and Mictomys, as in the present work ; whereas Microtus included eleven subgenera, namely, Eothenomys, Anteliomys, Lagurus, Alticola, Hyperacrius, Pedomys, Phaiomys, Pitymys, Chilotus, Microtus, Arvicola, andNeofiber, all of which are now accorded full generic rank. Ellobius, here included in the Microtinse, was not regarded as a member of the subfamily. Miller's work both stimulated and facilitated further research, so that in current literature more than fifty genera and subgenera (including fossil forms) are recognized. But some of these are not valid, and here only thirty-one genera and subgenera are described. The most important recent addition to our knowledge con- cerns the remarkable genus Prometheomys described in 1901 by Satunin. Vinogradov has prepared a valuable and beautifully illustrated account of some material now in Leningrad ; that paper, to which I am greatly indebted, will be published elsewhere. My knowledge of the genus rests, however, upon an even more solid foundation, thanks to the Hon. Ivor Montagu, who visited the Caucasus in the interests of the British Museum last autumn. There he obtained a magnificent series of specimens representing all the stages of growth of this rare animal; this material has enabled me to reach a definite conclusion as to the relationship of the genus and substantially to confirm Vinogradov's results. As regards living species my work has been greatly lessened by the labours of Miller and Thomas in the Old World ; and by those of Bailey, HoUister and many others in America. The collection of North American Microtinaj in the British Museum, although sufficient to enable one to appreciate generic characters and those of some of the leading species, is far too small and incomplete to warrant any attempt being made to give an inde- pendent and new account of American species and subspecies. But revisions of the North American species of Evotomys and ' North American Fauna, No. 12, 1896. 4 MICROTIN^ Microtus (in the wide sense understood by Miller in 1896) by Vernon Bailey ^ and of Ondatra by Hollister ^ have been pubUshed by the United States Biological Survey; in order to make the present book as compendious as possible free use has been made of this admirable work. Knowledge of fossil Microtinse has grown slowly since the days of Buckland and Cuvier, who seem to have been the first to pay attention to this part of the subject. In this country palseonto- logical work on the group has been done chiefly by Owen, Sanford, Blackmore and Alston, E. T. Newton, Forsyth Major and the present writer; on the Continent we are indebted chiefly to Pomel, Hensel, Forsyth Major, Nehring, Woldrich, and Mehely. In America fossil remains have attracted less attention : Cope and Hollister are the chief writers who have dealt with them hitherto. Mention must be made of the work of Forsyth Major and Winge, two distinguished men who paid much attention to the Microtinse; their respective contributions to the special hterature of the group are among the most valuable we possess. In the writings of Forsyth Major we find the one theory that, in my opinion, fits all the facts and accounts satisfactorily for the evolu- tion of rodent molars. In those of Winge is a masterly review of all the mammalian orders, together with a lucid exposition of the principles which should be used as a guide when investigating the relationships of mammals. One may differ from Winge on almost every point of detail and yet find inspiration and guidance in his work. Every mammal is the product of two distinct and sometimes conflicting forces ; a compound of relatively essential characters, fixed for the time being in each group by inheritance, and of more or less plastic characters which yield like potter's clay to the thumb of stern necessity. It is the special use that a mammal makes of its various organs that results eventually in a more or less perfect adaptation of form and structure to particular func- tions, no matter whether the special use is called into being by tempting opportunity or by the compelUng stress of circumstances. Use and habit, all that goes to make environment in the widest sense, have thus made species what they are; no character is absolutely beyond the reach of external influences, although in practice some may never be reached. That characters so acquired become in the course of generations intensified, customary, and at last the normal heritable attributes of a species is the lesson taught by every scrap of philosophical palseontological work upon the Mammalia that has been done since the day when Kowalevsky published the celebrated introduction to his memoir on Anthra- ^ Bailey, " Revision of the American Voles of the genus Evolomys," Proc. Biol. Soc. Washmgton, 11, pp. 113-138, 1897. Bailey, " Revision of American Voles of the genus Microtus" N. Amer. Fauna, No. 17, 1900. 2 Hollister, " A Systematic Synopsis of the Musk Rats," N. Amer. Fauna, No. 32, 1911. EXTERNAL CHARACTERS 5 cotherium; it is the lesson turned to practical use by Winge in systematic work. If the more or less substantial mask of specialization be stripped off from each species one finds the primitive core of each animal underneath; if the primitive characters so found be used as the bases of comparison there is no difficulty either in arranging species, genera, and families in natural order, or in conceiving what the essential characters of the ancestor common to any given group must have been. That is what I have attempted to do in this book. The stripping process is, however, by no means easy, and it reveals many a disconcerting gap in our knowledge. c. General Characters of the Microtin^. External Characters. In general outward form the Microtina3 differ but little from ordinary rats and mice. All are more or less evidently modified for burrowing, and a few, in addition to their fossorial peculiarities, show conspicuous adaptation for aquatic habits. The members of the subfamily thus display a somewhat striking uniformity in outward appearance, and lack that variety of shape which is so characteristic of the Cricetinee and Murinae. In relation to their size voles and lemmings are robust, thickset animals with broad, more or less flattened heads and short, bluntly rounded muzzles. The eyes and ears are small and in some genera are very greatly reduced. The limbs are moderately long, muscular and powerful; but they are hidden to a great extent in the general integument of the trunk, a circumstance which gives the Microtince a characteristic short-legged appearance. The hands and feet have each five digits armed with claws which differ considerably in size and form, according to the habits, in different genera ; the thumb is always greatly reduced and bears either a small claw or else a flattened nail. When least modified the palms bear five, the soles six well-developed pads ; but in the more highly speciaUzed forms some or all of these pads may lose their functional importance and tend to disappear. Typically the palms and soles are naked between the pads, or but scantily and incompletely clothed with hair ; but in the more specialized forms they may become either completely naked, as in the aquatic genus Ondatra, or densely covered with hair, as in the boreal genus Dicrostonyx ; in both these genera the modification is accompanied by reduction of the palmar and plantar tubercles. The upper surfaces of the hands and feet are always well clothed, but the lower surfaces of the digits are u-sually naked, with the skin thrown into scaly annulations by transverse folds. The tail is never very long, but may reach a length about two- thirds that of the head and body ; in several genera it is quite short and in some is reduced to a mere vestige considerably shorter 6 MICROTIN^ than the hind-foot. The skin of the tail is scaly, the scales usually forming rather well-marked annulations; it is more or less well clothed with stiff hairs, which may or may not conceal the scales ; frequently a distinct terminal pencil of variable length and thickness is developed. The fur tends to be soft and dense in all the Microtinse; in those most highly speciahzed for fossorial life it is very short, fine, uniform, silky, and mole-like in texture; whereas in the aquatic forms the contrast between the silky, dense under fur and the longer, stifier, and more lustrous contour-hair is intensified. The normal mammary formula in the group is 2 — 2 = 8, there being two pectoral and two inguinal pairs of mammse in the females. In many genera one or both pairs of pectoral mammae become functionless and disappear, the formula being reduced to 1 — 2 = 6 or 0—2 = 4; exceptionally the inguinal mammae suffer reduction, so that the formulae 2 — 0 = 4, 2 — 1 = 6, and 1 — 1 = 4 are not unknown within the subfamily. In many genera specially developed glands are present upon the flanks or hips in adult males and sometimes in both sexes. Special perineal glands, secreting a powerful musk, occur in the genus Ondatra. Skull. The skull of the Microti nae is always of firm and often of relatively massive construction, the sagittal sutures between the paired frontal, premaxillary, maxillary, and palatine bones generally fusing and disappearing either before or shortly after birth. Normally the skull is characterized by the shortness of the rostrum (in dorsal view), and by the forward position of the orbit, which is always anterior to a vertical plane touching the front edges of the anterior molars. The form of the outer wall of the infraorbital canal is also characteristic ; it is a stout, more or less vertical plate of bone (the " masseteric plate " or " descending ramus of the maxillary root of the zygoma "), which is always placed more or less transversely or obliquely to the long axis of the skull; the front edge is slightly emarginate or undercut and never projects in advance of the front border of the superior ramus of the maxillary root of the zygoma. The infraorbital canal itself is formed essentially as in other Muridae, its upper portion being widened for the transmission of a slip from the masseter medialis muscle, its lower portion, narrow and slit-like, serving for the passage of the facial or infraorbital branch of the superior maxillary or second division of the fifth nerve and the accompanying vessels; in Ellobius, however, the lower part of the canal is reduced and closed, but the upper 25ortion, in compensation, is somewhat more spacious than usual. The zygomatic arches are strongly built and more or less widely bowed laterally; each is greatly strengthened by the oblique position and peculiar shape of its lower maxillary root (the " masseteric plate " described above), wbicb forms a great flying buttress for the support of the fore-part of the arch ; the superior maxillary root, forming the roof of the infraorbital canal, is very short. The distal end of each maxillary zygomatic process is expanded into a broad fork which receives the anterior end and much of the lower border of the jugal; this bone is always very short and is confined to the central portion of the arch, bridging the small interval between the maxillary and squamosal zygo- matic processes. The squamosal root of the zygoma is relatively weak. The nasals are relatively short, often ending anteriorly well behind the front faces of the incisors, and never conspicuously in advance of them ; posteriorly they terminate usually either a little in front of or on a level with the anterior margin of the orbit. The interorbital region is always clearly defined, but in its degree of constriction it differs considerably in different genera and species, as well as in different stages of growth in one and the same individual, tending as a rule to become narrower with age. In many forms the temporal ridges, which always extend far forwards, approach each other and in adults fuse to form a median crest in the interorbital region; in other forms the ridges though salient remain more or less widely separated by a median sulcus ; in many genera the ridges are feeble and both the interorbital region and the interval between the ridges in that region remain wide and flat even in old age. On the roof of the braincase the temporal ridges usually diverge from the hinder part of the interorbital region to points above and nearly in the same transverse plane as the glenoid articulations ; thence the ridges converge ■ slowly to the hinder part of the parietal region and then again curve outwards, passing the extremities of the interparietal to join the front face of the occipital crest. Throughout their course, after leaving the frontals, these ridges tend to follow the upper edge of each squamosal, but they impinge upon the parietals anteriorly and cross the lateral wing of each parietal in the post-glenoid region. In some genera (Ondatra, Prometheomys, and Ellohius) the temporal ridges are closely approximated throughout in the later stages of growth; and in Prometheomys they form a salient sagittal crest which, in old age, extends from the interorbital region to the occiput. The squamosals are always largely developed, forming a great deal of the sides of the braincase. In this subfamily no post- orbital processes are ever formed by the frontals ; but very charac- teristic post-orbital crests are developed upon the squamosals for the attachment of a tendinous portion of the anterior part of the temporal muscle on each side; in some genera these crests are represented by prominent peg-like processes. In genera in which the temporal ridges form an interorbital crest, the squamosals frequently show a tendency to approach each other anteriorly, MICROTINjE Fig. 1. — Arvicola amphibins Linnscus. Dorsal views of skulls in different stages of growth. Condylo-basal length, Stage I, 25-1 mm., St. II, 31-2 mm., St. Ill, 35-5 mm. Figs. I and 2 illustrate the changes, which accompany growth, in the relative proportions of the cranial and facial parts of the skull, the progres- sive constriction of the interorbital region, gradual development, approxi- mation and fusion of the temporal ridges, increasing salience of the post- Fig. 2. — Ariicola ampliihius Linnaeus. Dorsal views of skulls in different stages of growth (continued). Condylo- hasal length, Stage IV, 37-8 mm., «t. V, 40-6 mm., St. VI, 44-3 mm. orbital crests, and gradual approximation of the squamosals on the fore-part of the braincase. Other changes in the rostrum, zygomata, interparietal and occiput are also shown. 10 MICROTIN^ Fig. 3. — Arvicola ampldhius Linnseus. Ventral views of skulls in different stages of growth. Stages I, II, and III (same specimens as shown in Fig. 1). In Stage I the rostrum is short and broad, the molars are relatively large, the palate is simple and level throughout, the pterygoid fossae are short, and the bullae are rounded. In the later stages the rostrum becomes progressively longer and narrower as the incisors increase in length, the 11 Fio. 4. — Arvicola amphihixis Linnoous. Ventral views of skulls in different stages of growth. Stages IV, V, and VI (same specimens as shown in Fig. 2). molars become relatively smaller, the hinder part of the palate is sculptured in bold relief, the pterygoid fossse increase in size, and the bullae develop large eustachian processes. 12 MICROTIN^ Fig. 5. — Arvicola amphibius Linnaeus, Lateral views of skulls in different stages of growth. Stages I and II (same sjiecimens as shown in Fig. 1). In Figs. 5-7 the progressive flattening of the dorsal surface, the changing direction of the occiput (inclined backwards in Stage J, forwards in Stage VI), the forward migration of the orbit (most marked in Stages V SKULL 13 Fig. C. — Arvicola atnphibius Linnaeus. Lateral views of skulls in different stages of growth. Stages III and IV (same specimens as sho\ra in Figs. 1 and 2). and VI), and the changes in the form of the auditory bull* are features worthy of attention. u MICKOTIN^ Fio. 7. — Arvicola amphibius Linnaeus. Lateral views of skulls in different stages of growth. Stages V and VI (same specimens as shown in Fig. 2). SKULL 15 encroaching upon the frontals to an unusual extent. In the post-glenoid region there is typically a large fenestration in each squamosal immediately above the auditory bulla; but. in some genera the fenestra is represented merely by a small foramen. The interparietal is usually large. In genera in which the tem})oral ridges sweep by and do not cross the lateral extremities of the bone its form is to a very large extent influenced by the degree in which those ridges are approximated posteriorly; as the ridges tend to become closer together with age, the inter- parietal becomes narrower in relation to its length. In genera in which the ridges traverse the interparietal that bone becomes very small, as in Ondatra and Prometheomys ; in the latter genus TEHaO Fig. 8. — Arvicola arwpliibius Linnaeus. Anterior and jjosterior views of skull with mandible in position : s.r. upper maxillary root of the zygoma ; i.r. lower maxillary root of the zygoma, forming the outer wall of the infraorbital canal or " masseteric plate " ; m. mastoid portion of periotic. the sagittal suture persists, even in old age, dividing the bone into two halves, a character that may be seen occasionally in other voles, e.g., Phaiomys. The anterior palatal foramina are usually moderately large, but in forms showing marked fossorial specialization of the skull they are much reduced in size. From the posterior end of each foramen a well-marked groove, the lateral palatal groove, runs back on each side of a median elevated tract and crosses the maxillo-palatine suture in the neighbourhood of m^ ; the groove may terminate here in one of the posterior palatal foramina ; but, if the 250stero-lateral bridges are well developed, it usually continues past the foramen, deepens behind it, and passes under (dorsal to) the postero-lateral bridge to effect, as a rule, a junction with the postero-lateral pit (" post-palatine fossa "). The posterior portion of the palate presents characters of 16 MICROTIN^ importance to the systematist. In some genera the palate terminates behind in a simple transverse shelf with or without a median process or " nasal spine " ; in these the postero-lateral pits lie at a deeper level and pass forwards under (dorsal to) the palatal shelf, and their inner borders do not form any externally visible connection with the shelf or with its " nasal spine " (e.g., Evo- tomys, Dicrostonyx, Lemmus). In other genera the " nasal spine " is elongated, is inclined dorsally instead of being horizontal, and is connected at its tip and sides with the inner borders of the postero- lateral pits, thus becoming converted into a sloping median septum between the pits. In the more primitive of these genera the median septum is short, broad, and ill defined, but in the more specialized ones it is long, narrow, and very sharply defined ; frequently the ventral surface of the septum is grooved, and some- times it is completely cleft. The floors of the postero-lateral pits are usually perforated by numerous small foramina which serve for the transmission of nerves and vessels to the soft palate and upper part of the pharynx ; but in Prometheoinys, in which the hinder part of the palate is generally speaking somewhat inter- mediate between the two types described above, the floor of each pit is occupied by a single very large foramen. The pterygoid fossae vary in size and depth from genus to genus, but the ectopterygoid plates are always well developed. When deepest the floors of the fossse lie at a level distinctly dorsal to the ventral surface of the basisphenoid and by means of a small transverse canal perforating the latter they are placed in com- munication with each other. To a large extent the size of the pterygoid fossae, the width of the choante, of the back part of the basisphenoid and of the fore-part of the basioccipital seem to be dependent upon the varying size of the auditory bullae. Some of the most characteristic features of the Microtine skull are shown in longitudinal vertical sections (Figs. 9 and 10). In other Muridae the palatine processes of the maxillary and palatine bones are thin ; but in Microtinae they are enormously thickened, probably in correlation with the unusually tall and robust crowns of the molars and with the powerful development of the jaw muscles. The hinder part of each palatine, i.e., the portion which forms the post-palatal pit, remains thin. The presphenoid, the basisphenoid and the fore-part of the basioccipital are con- siderably thickened vertically. In the floor of the braincase (Fig. 11) the presphenoid is of unusual length, in correlation with the great length of the tooth-rows ; but owing to the intrusion of the alveolar capsules of m^ and m^ into the sphenorbital fissure, this bone and the fore-part of the basisphenoid are greatly com- pressed laterally. The sphenorbital fissure occupies a great area in the cranial floor, but in the more hypsodont genera its outlets are greatly straitened by the tooth capsules, which divide the fissure into two parts, an inner and an outer ; the inner division between the tooth capsules and the presphenoid transmits the SKULL 17 first and second branches of the fifth nerve ; the outer division between the tooth capsules and the ahsphenoid serves for the passage of the internal maxillary artery, which often grooves the outer side of the alveolar capsule of ?u^. The alisphenoids are very large ; of each the ascending process forms a large part of the anterior wall of the braincase, rising in the orbit nearly to the level of the post- orbital process of the squamosal and there articulating with the frontal ; the descending process forms an ectopterygoid plate, greatly developed in all Microtinae, which articulates with the squamosal and auditory bulla behind, and bridges the foramen ovale to serve as a buttress supporting the hinder end of the alveolar portion of the maxilla. The horizontal plate is anchylosed with the basisphenoid as usual ; it is pierced at its base by a foramen for some of the cerebral vessels in front of the very shallow pituitary fossa, and externally by the large foramen ovale. Just below the front edge of the foramen ovale and external to the ectopterygoid plate is a foramen by which the internal maxillary artery enters the outer division of the sphenorbital fissure. There is no distinct foramen rotun- dum, the second branch of the fifth nerve emerging with the first through the sphenorbital fissure. The orbito-sphenoids are small, are completely anchylosed with the presphenoid, and are pierced by the small optic foramina, as usual. Posteriorly the basisphenoid is often notched for the passage of the internal carotid artery ; in Prometheomijs the notch is represented on each side by an oblique tunnel, the mouth of which is a conspicuous feature in the floor of the braincase; small passages from the anterior wall of this tunnel place it in communication with the vascular passages through the base of the alispheuoid and also with the transverse canal, which is constantly present in the Microtinte and gives passage to a vein that passes from one pterygoid fossa to another through the body of the basisphenoid. The auditory bullae are always well developed. When least modified they have thin papery walls; when more specialized their walls are strengthened by bony threads developed in folds of the mucous membrane lining the cavity of the middle ear; and in their most modified condition the walls are formed by a compact mass of spongy bone, and the teg men tympani and mastoid portions are considerably inflated. The external meatus becomes tubular in aquatic forms, and is often much straitened in fossorial genera. The stapedial artery is often enclosed in a bony tube which passes through the stapes. The mandible always possesses distinct coronoid and angular processes, although the latter are reduced in some genera. The horizontal ramus is characteristically stout, and is considerably thickened for the accommodation of the hypsodont cheek-teeth. At the symphysis the mandibular rami are united, as in other Muridse, merely by ligament. 18 MICROTIN^ Dentition. As usual in Muridse the dentition of the Microtinse consists of sixteen teeth, namely a pair of incisors and three pairs of cheek- teeth or molars in each jaw above and below, usually expressed by the formula i\ m| = 16. It is, however, questionable whether the three cheek-teeth are in fact the homologues of the teeth called m\, to|, and mf in other placentals (see p. 124). As in all other Rodents the incisors are persistently growing Fig. 9. — Vertical longitudinal sections of skulls of Muridae (enlarged). a. Raltus raltus Linnaeus; 6. Neotoma sp. (See p. 16.) teeth. The upper incisors are strongly curved as compared with those of the lower jaw and form larger segments of relatively small circles, whereas their opponents invariably form smaller segments of relatively large circles. Posteriorly the alveolar sheath of each upper incisor passes backwards into the maxillary bone, where it usually terminates just in front of the alveolus of m^; in Ellohius, however, it pushes its way back on the inner or lingual side of the molar roots to terminate in the hinder part of the maxilla immediately dorsal to the palatal surface of the bone, the termination of the tooth capsule being often marked by a fenestra- tion of the maxilla. Each lower incisor passes backwards through the mandibular ramus. In the Lemmings this tooth lies on the TEETH 19 lingual side of the molars throughout, and its alveolar capsule terminates behind at a point opposite the alveolus of 7n.^. In the Voles the tooth is much longer; it passes backwards on the lingual side of dij and m^, crosses to the labial side of the jaw A.B. Bo. Fio. 10. — Vertical longitudinal sections of skull of Arvicola amphibius Linna-us (enlarged). I. P. interparietal ; Sq. squamosal ; //. optic foramen ; /. olfactory chamber ; V. vomer; As. alisphenoid; Me. mesethmoid; B.s. basisplienoid ; c.l. transverse canal; P. palatine; P.s. presphenoid; M. maxilla; I .a. alveolus of incisor. I'b. turbinals; F. frontal; P. parietal; /. supratympanic fenestra of squamosal; P.m. premaxilla ; w^ m*. alveolar capsules; P<. pterygoid ; A.B. auditory bulla; Bo. basioccipital. between ^Hj and m^, and ascends into the condylar process to a greater or less height, its termination often producing a well- marked hump on the outer surface. In structure the incisors are quite typical. Each is developed from a persistent pulj) lodged in the widely open pulp-cavity at its base. The front face of the tooth is formed by a thick plate of hard enamel which thins out and ends off on the lateral surfaces ; 20 MICROTIN^ the core of the tooth, exposed on its posterior face, is formed of relatively soft dentine. In transverse section the teeth are usually wider than deep ; the enamel is frequently stained by a yellow pigment; the front face of each upper incisor may or may l.af. Fio. 11. -Skull of Arvicola amphibius Linnseus, with top removed to expose the floor of the braincase. P.m. premaxilla; N. nasal; L. lachrymal; M. maxilla; F. frontal; •/. jugal; O.s. orbitosphenoid ; P.s. presphenoid; A.s. alisphenoid; Sq. squamosal; B.s. basisphenoid ; A.B. periotic; B.o. basioccipital. I.o.f. infraorbital canal; I. olfactory chamber; //. optic foramen; m^, m^. alveolar capsules; a.m.i. groove for internal maxillary artery; F.1.2. exit for first and second branches of fifth nerve ; F*. foramen ovale ; F.l.m. foram,en lacerum medius. not be traversed by a longitudinal groove. The teeth differ considerably from genus to genus, and the significance of some of the differences is discussed below. The cheek-teeth are highly characteristic. In all Microtinse they are tall-crowned or hypsodont. In some genera they are of limited growth, closing their pulp-cavities and cement-spaces below, developing roots, and wearing out in old age. But in the TEETH 21 majority of the genera they are persistently growing teeth Jike the incisors, in which development at their bases from ever-active dentinal pulps and enamel organs compensates throughout life for Fig. 12. — Arvicola ampJiibius Linnseus. Skull and left mandibular ramus dissected to show the alveolar courses of the teeth. the continuous loss of substance which takes place at the wearing surface of each tooth. The worn surfaces of these teeth display a peculiar pattern of triangles and transverse loops, produced by the truncation by wear of the ends of the tall columns or prisms of which each 22 MICROTIN-E crown appears to be composed, and giving rise to a number of salient angles and re-entrant folds along the inner and outer borders of the tooth. Each triangle or loop is formed by dentine, surrounding an inner core of relatively soft " osteodentine," and bounded externally by a sheet of relatively hard enamel. The front end of each upper molar and the hinder end of each lower molar is formed by a more or less crescentic or pyriform transverse loop, the inner and outer extremities of which form the first inner and first outer salient angles. The triangles behind or in front of this transverse loop are arranged in two series, an inner and an outer, the members of which alternate with each other more or less distinctly and regiilarly. The apex of each triangle forms a salient angle which is separated from its neighbours by a re- entrant fold or cement space, which in many genera is partly filled with cement. At the hinder part of m^ and the fore-part of «rj the triangles are succeeded by a complex structure, the " pos- terior loop " in m^, the " anterior loop " in m^ ; these terminal loops are often of great systematic importance. In describing the teeth it is customary to enumerate the salient angles and re-entrant folds from before backwards in upper molars, and from behind forwards in lower molars, the first salient angle on each side being formed by the transverse loop. The differences in the number, form, and relative size of the triangles and salient angles, the degree to which the dentinal spaces are open to or closed off from each other, the greater or less complexity of the anterior loop in m^ and of the posterior loop in w?, the distribution and nature of the enamel sheet in different parts of the periphery, the presence or absence of cement, and above all the circumstance whether the cheek-teeth are of persistent or of limited growth, i.e., rooted or rootless; — all these, when used with discretion, afford excellent characters for the distinction of genera and species. But in using them for such purposes it must always be borne in mind that the pattern is often subject to considerable variation, not only in different individuals, but in different stages of wear in the same individual ; the old idea that " prismatic teeth " present the same pattern throughout life is erroneous. These and many other points are discussed below in the section dealing with the evolution of the teeth (p. 102). Jaw Muscles. It is not possible to give a full account of the myology of the Microtinse in the present work, but the following notes on the more important muscles of the jaws, based chiefly upon several dis- sections of Arvicola amphibius, will give a fair idea of the general arrangement found in the group. The most important muscles are the temporalis and the masseter (Figs. 13 and 14). These have played a great part in moulding the outward form of the skull; the effects of their JAW MUSCLES 23 powerful influence upon the bones to which they are attached are seen not only in passing from genus to genus, but in tracing the life-history of the individual from youth to old age. Masseter lateralis. This muscle consists of two distinct portions ; one, anterior and superficial, partly concealing the other, posterior and deep . The anterior portion arises by a strong tendon from the lower border of the maxilla just in front of and below the mouth of the infraorbital canal. Becoming fleshy and wider, its fibres pass backwards obliquely to their insertion which occupies the thick- ened lower border of the angular process of the mandible. Along the inner edge of that border the fibres meet the ends of those of the ^pterygoid internus muscle. Owing to the oblique, almost horizontal, course of its fibres this portion of the masseter is that principally concerned in drawing the jaw forwards when the animal is gnawing. The posterior portion rises by fleshy fibres from the whole outer surface of the outer wall of the infraorbital canal and from the lower border of the zygomatic arch throughout its entire length. The muscle is inserted into the whole length of the masseteric crest of the mandible from a point below m^ to the end of the angular process behind. The anterior fibres become tendinous towards their insertion, and they have a more nearly vertical direction. The fibres become more and more oblique posteriorly, the hindermost winding round the bases of the angular and condylar processes to be inserted upon the inner surfaces of the upper part of the angular process and the hinder base of the condylar process. This portion of the masseter is very powerful, drawing the jaw forwards and upwards. Masseter medialis. This is a comparatively feeble muscle arising by fleshy fibres from the inner surface of the zygomatic arch, the inner surface of the outer wall of the infraorbital canal, and by a small slip, which, passing through the infraorbital canal, arises from the side of the maxilla in the prezygomatic fossa. The anterior fibres, from the infraorbital canal and floor of the orbit, pass downwards and backwards and are inserted by tendon into the fore-part of the masseteric crest under cover of the posterior portion of the masseter lateralis. The fibres from the zygomatic arch pass vertically downwards to their insertion by tendon into a groove which rises from the crista masseterica below and ascends parallel with the front border of the coronoid process to a point a little below the base of the sigmoid notch ; thence the line of insertion passes backwards to the alveolar protuberance of the lower incisor and finally ascends the condylar process to a point a little below the head. The small posterior portion of the muscle arising from the squamosal root of the zygoma and inserted into the fine around the sigmoid notch, is usually separable from the chief or anterior portion, a branch of the facial nerve passing out between them. The masseter medialis assists the temporalis in 24 MICROTIN^ closing the jaws, and its most anterior fibres pull the lower jaw forwards as well as upwards. Temporalis. This muscle although varying considerably in size and in its precise relation to the braincase in Microtinse M. M. a. Fig. 13. — Arvicola amphibius Linnaeus. Jaw muscles : M.L.a. anterior or superficial portion of masseter lateralis ; M.L.p. posterior or deep j)ortion of masseter lateralis; M.M.a. and M.M.p. anterior and posterior portions of masseter medialis; T. temporalis. is always well developed. The temporal fossa extends from the occipital crest to a point more or less far forwards in the inter- orbital region. The muscle is covered by a fascia which extends from the temporal ridge, marking its upper limit of origin, to the JAW MUSCLES 25 upper border of the jugal and to the lower edge of the temporal fossa in the post-glenoid region. The temporalis is formed chiefly by a great fan of fleshy fibres which arise from the floor of the temporal fossa ; the more superficial fibres take origin from the Fig. 14. — Arvicola amphibius Linnaeus. Jaw muscles (continued) : letters as in Fig. 13. under surface of the temporal fascia. The fibres converge to their insertion into the tip and inner face of the coronoid process. The anterior and central portions of the muscle are particularly strong ; and some of the deeper fibres arising from the squamosal upon the anterior shoulder of the braincase have a tendinous origin, the tendon producing the post-orbital crest or process of the squamosal which is so characteristic of the subfamily. 26 MICROTIN^ Parsons ^ describes and figures the " parietal portion " of the temporalis in the Water Vole as being continuous with what is described above as the anterior portion of masseter medialis. The foregoing description is based chiefly upon several dissections of the jaw muscles in the same species ; I have never found any portion of the temporalis going to an insertion external to the coronoid process (apart from the fibres that envelop the tip of the process) and my experience in this respect seems to agree with that of Tullberg.2 Pterygoid internus. This is a short, thick muscle passing obliquely backwards, outwards, and downwards from its origin in the pterygoid fossa to its insertion upon the inner face of the mandibular angular process. Pterygoid externus. A small muscle arising from the ectopterygoid plate and inserted upon the inner face of the condylar process of the mandible. DiGASTRicus. This muscle arises from the paroccipital process and is inserted into the lower border of the mandible towards the symphysis. Anteriorly its belly is in close contact with that of its fellow of the opposite side. Its function is to retract and depress the mandible. d. The Evolution and Status of the Subfamily Microtin^, AND THE Interrelationships of the Genera. There can be no doubt that the Microtinse have descended from a primitive Myomorphous stock which was also ancestral to all the other groups of Muridse. Among Muridse the sub- family is sharply defined by its cranial and dental characters. Of these the most prominent are the firm construction of the skull, shortened rostrum, forwardly placed orbits, peculiarly formed " masseteric " plates, the presence of post-orbital squamosal crests or processes, the thickened palatal processes of the maxil- laries and palatines, and the hypsodont prismatic cheek-teeth. The molars of certain Cricetinse {e.g., the North American genera Sigmodon and Neotoma) show a strong superficial resemblance to the cheek-teeth of Microtinae; but detailed study shows that similar tooth-patterns have been evolved in somewhat different ways in the two subfamilies. In the remarkable Asiatic genus Myospalax (" Siphneus ") the cheek-teeth are rootless and closely resemble in pattern those of the typical lemmings among Micro- tinse. But although the skull (Figs. 15-17) is highly speciaHzed for fossorial habits, it and the jaw muscles resemble those of the Cricetinse in retaining essential features similar to those found in the more primitive of the non-Microtine Muridse generally, and differing widely from those characteristic of the Microtinse. Among the most primitive Nesomyinse (e.g., Nesomys), now 1 Paesons, P.Z.S., 1896, p. 160. 2 TuLLBEEG, Ueber d. System der Nagetiere, Taf. xiv, figs. 17 and 20. EVOLUTION 27 Fig. 15. — Myospalax fontanus Thomas. Skull of adult, dorsal view, enlarged ; the smaller figure natural size (B.M., No. 9.1.1.203, Shansi). 28 MICROTIN^ confined to Madagascar, we find cheek-teeth which, although very- different from those now characteristic of voles, present an arrange- ment of the crown tubercles similar to that which, in a less reduced condition, probably characterized the molars of the ancestor of the Microtinse. In one genus of this Malagasy group, Brachytarsomys, which has adopted a fossorial vole-like mode of living, the molars, although low-crowned and rooted, have been simplified trans- versely in such a way that they show all the broader features Fig. 16. — Myospalax fonlanus Thomas. Skull of adult, ventral view, enlarged. observable in the cheek-teeth of voles. It is of great interest to observe that in this genus (and in this genus alone) the jaw muscles have developed exactly as in the highest voles; and the skull of Brachytarsomys, although very primitive in many respects, yet makes an undeniable approach towards that of the Microtinse in all those features which in voles depend upon the special development of the anterior part of the temporalis and upon the peculiarities in the arrangement of the masseter lateralis muscles. The parallel, indeed, is so close that it may be necessary later on to transfer Brachytarsomys to the Microtinse. The facts EVOLUTION 29 mentioned suggest that the dental, cranial, and myological characters are inseparably linked with each other both in Brachy- tarsomys aud in the voles. The problems connected with the evolution of the patterns of the cheek-teeth and their bearing upon relationships are more fully discussed below at p. 102. The Microtinfe owe their development, special characters, and survival in the face of keen competition, chiefly to the fact that they have acquired the power of subsisting upon coarser, tougher, Fig . 1 7 . — Myospalax fontanus Thomas. Skull of adult, lateral view, enlarged. less inviting and less nutritious vegetable substances than those devoured by their more generalized relatives and rivals. For the most part, too, the Microtinse have become earth-bound, burrowing creatures, and their fossorial habits have played a great part in moulding both the outward form and the internal structure of most of the living genera. The immediate ancestor of the Microtinse must have been a generalized Murine with moderately large eyes and ears, long tail, normal feet and hands, and normal fur. In these respects it cannot have shown a greater degree of fossorial specialization than that e.xhibited in Brachytarsomys. Its skull must have resembled 30 MICROTIN^ that of the more primitive Muridae in its general lightness and delicacy of construction, in the persistence of the median sutures, the flatness of the palate, the wide separation of the temporal Fig. 18. — Brachytarsomys albicauda Gunther. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skuU in dorsal view, natural size. ridges, and in many other respects. The upper incisors when unworn retained distinct traces of their original cuspidate crowns (see pp. 99-101). The lower incisors were wholly lingual or inferior to the molars in position; they were very short, so that they did not extend further back than to a point opposite the EVOLUTION 31 hinder edge of the alveolus of mj. The cheek-teeth, as explained fully on pp. 102-119, must have been low-crowned, rooted, and niultitubercular ; their crowns were remarkably complex, m, consisting of about twenty-one tubercles; each of the other teeth was composed of about fifteen tubercles, the tubercles being arranged in three longitudinal rows both in upper and lower molars. As in other primitive Muridse the food was bruised and crushed, the motion of the lower jaw during mastication being transverse or oblique instead of longitudinal in direction. From such a form all known Muridse living or extinct may well have descended ; but in this work we have merely to trace the modi- fications which have led to the evolution of the Microtinse from this primitive type. Representatives of the primitive stock just described appear at different times to have discovered the nutritive value of the coarser vegetable substances such as moss, grass, tough leaves, bark and roots, and by degrees they substituted such unattrac- tive provender for the softer and more succulent fruits, berries, nuts, tender foliage and green shoots eaten by their ancestors and less enterprising rivals. Leaving the dainties to others, they thus tapped vast and never-failing food supplies which have in the course of time enabled their descendants, the Microtinse, to colonize the Holarctic region more thoroughly and more completely than any of the related groups. This gradual change in diet induced corresponding gradual changes in the dentition, ahmentary canal, and all the related organs. Gnawing, that fundamental habit of Rodentia, lost some of its primitive importance, since the extraction of kernels from hard shells ceased to be one of the chief operations in feeding, and the incisors, free to develop in other directions, were used sometimes as digging instruments, sometimes as forceps for the extraction of seeds, etc., special uses which have led to various modifications in the form of these teeth (see pp. 99-101). The new food made great and ever-increasing demands upon the cheek- teeth, and old normal methods of mastication had slowly to be changed. Bruising and crushing no longer sufficed for the reduction of the food, and were gradually replaced by shearing and slicing. Low-crowned tuberculate teeth, admirably adapted for the earher use, had to be transformed into tall-crowned, prismatic structures fitted for the new purpose. Mere pressure between the tooth-rows, accompanied by a transverse or oblique rocking motion of the lower jaw, had to be replaced by a powerful gliding stroke from behind forwards. The motion imparted to the jaw by the masseter lateralis muscle, used by other Muridae chiefly in gnawing, had now to become the chief motion in masti- cation, and the anterior part of the temporalis muscle became specially developed as an auxiliary to the masseter complex for this purpose. The uneven surfaces of the primitive tubercular teeth had at first to be worn flat in order to permit the lower jaw 32 MICROTIN^ to glide forwards when the molars were pressed tightly together ; but as the forward stroke became firmly established the crown tubercles gradually ceased to be functional parts of the crowns. The enamel, primitively rather thick, equally developed, and continuous in all parts of the periphery of the crown, became differentiated into thick and thin portions, remaining thick where required, becoming thin or disappearing altogether in situations where it was no longer useful or impeded the stroke of the teeth. In this way the enamel has been re-arranged in upper and lower molars to form a series of appropriately curved cutting blades which shear with each other effectively as the lower jaw is pulled forwards and upwards by the muscles. The apical enamel has atrophied and disappeared to a very large extent, so that the molars of many voles now present as soon as they cut the gum a flat surface upon which hard enamel and soft dentine are already exposed in effective alternation. Step by step, as the food has increased in harshness, the rate at which the substance of the teeth is wasted by attrition has become more rapid; in compensation the crowns of the teeth have become progressively taller, their dentinal pulps and enamel organs more vigorous and more con- tinuously active; and the loss of that vigour and activity has been postponed to a later and later moment in the life of the individual from generation to generation, until at last in the highest Microtinse the molars like the incisors have acquired the power of persistent growth. Other direct effects of the change of diet are modifications in the alimentary canal. The food of the Microtinse is richer in cellulose than that of less specialized Muridte and the process of digestion is chemically somewhat different. This has led, as in other rodents which subsist upon similar food-stuffs, to marked enlargement of the caecum and to enlargement and complication of the large intestine.^ The increased height of the molar crowns and their robust proportions have necessarily led to the enlargement of the alveolar capsules in which the teeth are developed and supported. In the upper jaw the capsules rise up as conspicuous swellings in the floors of the infraorbital canal and nasal chamber (m^) and in the floors of the orbit and sphenorbital fissure (m^ and m^) ; in the lower jaw they fill the body of each horizontal ramus and with the contiguous shaft of the lower incisor impart to the mandible a characteristic robustness of form. In those less specialized Microtinse in which the molar teeth still develop roots in old age, senility is marked by the gradual subsidence and collapse of the alveolar capsules ; in these forms, therefore, the alveolar portions of the jaws revert in old age to the condition seen throughout life in primitive Muridse. But in all Microtinae the great size of the alveolar capsules, whether a permanent or a more or less temporary feature, has produced important modifications of the surrounding ^ TuLLBERG, Ueber d. System der Nagetiere, p. 443. EVOLUTION 33 parts. The palatal processes of the maxillaries and the palatines have been deepened vertically ; the median sutures of these bones, open throughout life in primitive IMuridse, have closed and disappeared ; the ectopterygoid plate of the alisphenoid supporting the hinder end of the maxilla has been greatly developed, and the presphenoid and the fore-part of the basisphenoid, owing to the intrusion of the alveolar capsules into the sphenorbital fissure and the consequent driving inwards of the first and second divisions of the great trigeminal nerve, have become laterally compressed and vertically deepened. Under the influence of the deep part of the masseter lateralis muscle the zygomatic process of the maxilla, particularly its lower root which forms the outer wall of the infraorbital canal, has become unusually strong and characteristically formed. The temporal muscle, and particularly its anterior portion, is greatly increased in size in all Microtina? ; it has made room for itself in the orbito-temporal fossa by driving the zygomatic arch outwards, by compressing laterally the interorbital region which in the highest forms becomes extremely narrow, and by driving the eye forwards. The latter organ, becoming less and less important as Microtinse become more strictly fossorial, is reduced in size and displaced, being lifted up by the molar capsules as well as pushed forwards by the temporal muscle. The squamosal bone, which gives origin to the greater part of the temporal muscle, has like this muscle risen in importance ; it shows a well-marked tendency to encroach upon the frontals anteriorly and upon the parietals and interparietal behind in many Microtine genera, and on the shoulder of the braincase it forms a more or less well-marked post-orbital process or crest for the origin of the tendinous portion of the temporalis. Partly in consequence of the increased develop- ment of the ectopterygoid plate, but partly owing also to increased size of the pterygoid internus muscle, the pterygoid fossa has been deepened on each side. In the mandible the muscular specializa- tions are reflected in the form of the coronoid and angular processes and in the strength of the crista masseterica ; where the inser- tions of the pterygoid internus and masseter lateralis muscles have become tendinous and concentrated, the angular process has suffered reduction. Fossorial specialization has led to the reduction of the eyes and external ears, shortening of the limbs and tail, broadening of the head and thickening of the body, and to many modifications of the hands, feet, and fur. In the skull it is betrayed by the flattening of the dorsal surface and, in extreme types, by the pro- jecting and straightened (proodont) incisors, shortened nasals, shallowed rostrum, and forwardly inclined occiput. The auditory bulla3 become specially developed for underground life ; the cavities of the middle ear, ma.stoid portion, and the swollen tegmen tympani are filled, in the most modified forms, with spongy tissue, and the canaliculus tympanicus becomes completely V.L. D 34 MICROTIN.E ossified. In the pelvis the pubic symphysis is characteristically shortened. Many of the interpretations of Microtine structure given in the foregoing paragraphs have already been made by Winge.^ I have, however, worked through and reflected upon the whole subject for myself many times, and have arrived in my own way at my own opinions and conclusions. These differ, in many important respects,^ considerably from those of my illustrious predecessor, to whom I am gratefully indebted for much kindness and instruction, and it is therefore necessary for me to shoulder responsibility for the whole. That the evolution of the Microtinse has proceeded along the lines described above and for the general reasons given, becomes evident not only when we review the known genera, but when we follow the post-natal development of the individual. Comparison of unworn tooth-germs with specimens in successive stages of wear confirms the views expressed above as to the character of the dentition of the ancestor and with regard to the manner in which the ancestral molars have been modified (see the special sections dealing with the dentition at pp. 99-124). Similarly the post- natal growth of the skull in various genera recapitulates many of the chief evolutionary processes which have resulted in the skull forms characteristic of the adult stages of growth in the most highly specialized genera. For example, new-born skulls of Dicrostonyx, Evotomys rufocanus, Arvicola, and Ondatra (Plate II) are all much alike and closely similar to the skull in adults of the most primitive Muridse. They naturally show the beginnings of the Microtine specialization ; but most of the essential characters of the group become more and more evident in later stages of growth. In two respects the Microtinse, judged by their most primitive forms, stand lower in the scale than all other Muridae. Firstly, their cheek-teeth retain more of the primitive longitudinal com- plexity than do those of any other subfamily. Secondly, in the lowest Microtinse, the Lemmi, the lower incisor is unusually short, not extending backwards beyond m^, and is wholly lingual to the molars in position. In all other respects, the Microtinse have been carried by the two-fold specialization described above to a level far above that attained by any of those members of other subfamilies of Muridse that have specialized in somewhat similar directions. ^ WiNGE, " Om Graeske Pattedyr," Vidensk. Medd. Naturhist. Foren. Kjobenhavn, 1881, pp. 36-50. " Jordfundne og nulevende Gnavere (Rodentia) fra Lagoa Santa," E. Mus. Lundii, 3, pp. 123-126, 1887. " Gronlands Pattedyr," Meddelelser om Gronland, 21, 1902, pp. 358-360, 382-386. " Danmarks Pattedyr," 1908, pp. 50-55, 68-79. " Pattedyr Slaegter," 2, 1924, pp. 37-45, 137-143. ^ As long ago as 1914 Winge and I were comparing our views on this subject, and he told me that I had got everything upside down. No doubt others will be of the same opinion to-day ! INTERRELATIONSHIPS OF GENERA 35 Among the Microtinaj (excluding Brachytarsomys) the Lemmi are at once the oldest and fundamentally the most primitive. The remarkable shortness of the lower incisor in all lemmings proves that the group diverged from the primitive Murine stock before that tooth had attained its full development; the development of the incisor in the lemmings was arrested by the decline of the gnawing habit sooner than in the voles. Other circumstances which point to the relatively greater antiquity of the lemmings are firstly, the fact that they are all peripheral forms with a relatively restricted and a waning distribution; secondly, that whereas the Microti are represented by numerous genera, including many in which the molars in adult stages of growth develop roots, the Lemmi are now represented only by a few genera in which, without exception, the cheek-teeth have acquired fully the power of persistent growth. In other words, only the most highly specialized Lemmi have survived; the more primitive forms with rooted molars having utterly disappeared before the com- petition of the newer and less primitive group of voles. DiCROSTONYX is fundamentally the most primitive of the lemmings, and is by its structure one of the most isolated genera of the subfamily. Its cheek-teeth, when quite unworn, have members of the three primitive longitudinal rows of tubercles distinctly developed; they retain, in adult stages of wear, many of the primitive terminal elements which are suppressed in most other genera, and their re-entrant folds lack cement. This primitive type has survived either by colonizing, or else by remain- ing in, the high north, where it has adapted itself to the rigorous conditions now obtaining in that region and to subsistence upon the unusually harsh and poor diet offered by that inhospitable land. The primitive characters are therefore hidden beneath a mask, the product of intense and two-fold specialization. The fur has become very thick and subject to both seasonal and geo- graphical variations of length, density, and colour, being shorter, thinner, and darker in summer, in warmer regions, and in the young ; longer, denser, and whiter or paler in winter, in the bleaker regions, and in old age. The peripheral parts are shortened and with- drawn to the shelter either of the dense fur or of the general integu- ment of the trunk ; thus the outer ears are reduced each to a mere fold of naked skin hidden in the fur ; the arms and legs protrude but little from the general covering of the body; the caudal vertebrae are shorter than the dense brush which envelops them. The hands and feet have become exceptionally broad, partly for digging, and partly for locomotion upon snowy wastes ; their palms and soles are densely clothed with crisp, curling fur, and the pads, having ceased to be of functional importance, are represented only by some feeble vestiges which can be found by clipping the hair away from the soles. The claws have become large and sharp, those of the third and fourth manual digits being highly modified for digging and subject to a remarkable and unique 36 MICROTIN^ seasonal change; with the approach of winter these two claws grow to an extraordinary size and develop a peculiar supplementary- ventral portion which sometimes surpasses the main part of the claw in length ; but with the return of spring this ventral portion is shed and the main claw is then worn down to normal length. The bones of the fore-arm, particularly the ulna, are greatly strengthened for the attachment of the powerful muscles which move the fossorial hand. In this genus the incisors have become slender, straightened, Fig. 19. — Dicrostonyx grcenlandicus Traill. Skull and left mandibular ramus dissected to show the alveolar courses of the teeth (enlarged). and more or less protruding; and they retain at the most but a very feeble vestigial trace of an anterior groove. The cheek- teeth, as in all other known lemmings, have become hypsodont, persistently growing, broad-crowned, with their enamel differ- entiated into thick and thin portions and reduced even to dis- appearance on those portions of the periphery of the crown where enamel is no longer of functional importance. The intrusion of the alveolar capsules of m^ and ?«^ into the sphen- orbital fissure has resulted in the lateral compression of the presphenoid, which is reduced to a slender rod in most species. The temporal muscles produce a salient ridge upon each side of INTEKRELATIONSHIPS OF GENERA 37 the braiucase and the attachment of their post-orbital tendons is marked by the prominent peg-Hke post-orbital process of each squamosal ; the anterior parts of these muscles produce an elevated Fig. 20. — Lcmmus lemtnus Linnaeus. Skull and left mandibular ramus dissected to show the alveolar courses of the teeth (enlarged). superciliary ridge on each side, but they are not strong enough as a rule to cause these ridges, even in old age, to fuse into an inter- orbital crest and the squamosal bones remain widely separated anteriorly. The cavity of the middle ear is partly filled with a dense sponge of bone. 38 MICROTIN^ The true lemmings, Synaptomys, Myopus and Lemmus, form an interesting group which, stands on a higher plane than Dicro- stonyx. In the shortness and wholly lingual course of the lower incisors the members of this group agree with Dicrostonyx and differ from the rest of the Microtinae, and when most modified the external form shows a specialization which in some respects is analogous to that seen in Dicrostonyx. But in other essential respects the group differs as widely from Dicrostonyx as it does from the voles. The persistently growing cheek-teeth are char- acterized by the reduction and loss of many of their primitive terminal elements, so that their patterns are as simple longitudi- nally as in the highest voles ; they are further characterized by the reduction of the outer salient angles in lower and of the inner salient angles in upper teeth ; by the tendency of the outer folds above, and the inner folds below to retain their original character of transverse valleys ; by their great breadth, by well-difierentiated enamel, and by the presence of cement in the re-entrant folds. In correlation with the more highly specialized molars, the anterior portions of the temporal muscles have apparently gained in strength and they produce more marked effects upon the portions of the skull directly under their influence ; thus the post-orbital tendon stimu- lates the squamosal on each side to develop a long, shelf -like, post- orbital crest ; the superciliary ridges fuse sooner or later to form an interorbital crest; the interorbital region itself is much constricted and the squamosals, in the adults of the highest forms, tend to meet anteriorly. Of the three genera of this group Synaptomys is the most primitive. Externally it differs little from such voles as Evotomys and wholly lacks the outward specialization for fossorial habits or a boreal habitat which characterizes the recent species of Lemmus ; the large ears and the, for a lemming, long and thinly clad tail are particularly noticeable. Other primitive character- istics are shown in the general lightness of the skull and but slight modification of the zygomatic arches, in the large anterior palatal foramina, in the opisthodont and grooved upper incisors, and in the persistence of vestiges of the middle row of tubercles in the molars of some species. On the other hand the posterior border of the palate, the floor of the braincase, and the auditory bullae show specializations which lead away from Myopus and Lemmus and afford parallels with those seen in some of the higher voles. Of the two subgenera, Mictomys and Synaptomys, into which the genus is divided, Mictomys is the more primitive in some respects. In it the lower incisor ends on the lingual side of the molars opposite the posterior end of //ij' whereas in Synaptomys it is a little longer, ending slightly behind the anterior end of m^; the mammary formula is 2 — 2 = 8 in Mictomys, instead of 1 — 2=6 as in Synaptomys. But in certain other features Mictomys is the more specialized. The strongly curved upper INTERRELATIONSHIPS OF GENERA 39 incisors are very broad and have the enamel brightly stained throughout in Synaftoimjs ; but in Mictomys they are more slender and the staining is weaker, disappearing towards the outer edge of the tooth ; in both subgenera each incisor bears a groove near its outer border in front, a memorial of the former cuspidate condition of these teeth. In correlation with the differ- ence in the upper incisors, the rostral part of the skull is stouter in Synaptomys than in Mictomys. The cheek-teeth have essen- tially the same patterns as those of Lemmus, but they present, frequently and persistently, traces of the median row of tubercles (y, z, etc.), and these traces indicate that although the tubercles in question have blended with the so-called principal cusps they are of large size and therefore of much importance. In m^ and m^ the enamel folds are a little more transverse in direction than in Lemmus and consequently they appear to be a little deeper ; in Mictomys these teeth show a vestige of cusp n, but in Synap- tomys the postero-internal corner of each tooth is rather more reduced than in Lemmus. In both subgenera m^ is peculiar, having the second transverse loop separated from the third principally by the very deep second outer fold, the first inner fold being very slightly developed. In the lower molars of Mictomys the outer folds are so slightly developed that the teeth have crenulate outer margins and no closed triangles; but in the lower molars of Synaptomys the outer folds are deep, closing off external triangles as in Lemmus. In general the molars of Mictomys are more primitive than those of Synaptomys, since they retain more of the primitive transverse arrangement, whereas those of Synaptomys have acquired more of that alternation of inner and outer elements which is perfected in Lemmus. In the genus Synaptomys the skull is less specialized than in Lemmus. The tips of the nasals project in front of the incisors. As in Lemmus the zygomatic arches are widest an- teriorly, but their expansion is not so great; the planes of their outer surfaces are nearly vertical instead of being con- vergent dorsally, and the upper borders of the jugals are much less boldly convex. The post-glenoid part of the braincase is not shortened ; the braincase is therefore longer and narrower, as well as less massive. Most of these characters are due to weaker temporal muscles and the absence of extreme fossorial specialization. The palate posteriorly is as in Microtus; in the subgenus Synaptomys it is much as in Microtus arvalis, but in Mictomys the posterior pits are extended at the expense of the median septum, which has become long and thin as in " Steno- cranius," and of the postero-lateral bridges, which are slender and incomplete. The molars are considerably narrower in proportion to their length than in Lemmus, but they diverge posteriorly as in that genus. The pterygoid fossae are a little longer than in Lemmus, but rather shallow, their floors being scarcely dorsal to the ventral surface of the basisphenoid. 40 MICROTIN^ In Myopus the external form is also vole-like, although more thickset than in Synaptomys. The skull and teeth agree in essen- tial respects with those of Lemmus, although the skull is dis- tinguished from that of Lemmus by its smaller size, lighter build, rather less expanded zygomata, anteriorly more widely separated squamosals, larger anterior palatal foramina, and more globular auditory bullse. In these various characters the genus shows itself to be more specialized than Synaptomys, although outwardly, and to some extent in cranial characters, more primitive than Lemmus. In Lembius the characters of the group reach their highest expression, the whole animal being greatly modified for fossorial habits. Outwardly this is shown by the extremely robust general form, small eyes, small ears hidden in the fur and destitute of meatal valves, large and broad hands and feet, enlarged fore-claws, peculiar large and flattened thumbnail, hairy palms and soles, with the pads reduced to functionless vestiges, and short, thick, and densely clothed tail. Under the influence of the powerful temporal muscles and of fossorial habits the skull has become massive, broad, and depressed, with very strongly built, abruptly and widely expanded zygomata; in the inter- orbital region the temporal ridges fuse to form a median crest; the squamosals form strong post-orbital crests and tend to approach each other anteriorly as age advances ; the braincase is square, its post-glenoid portion much shortened, and it seems to have been pushed forwards, encroaching upon and reducing the size of the temporal portions of the orbito-temporal vacuities ; the cheek-tooth rows diverge rapidly behind ; the palate termin- ates simply behind, the posterior median sloping septum being represented merely by a short, free, spinous process ; the ptery- goid fossEe are very short and deep. The auditory bullae are less inflated than in the related genera, but the cavity of the middle ear is filled with a dense sponge of bone. The incisor teeth are rather slender, the upper ones rather strongly curved and with- out well-marked anterior grooves. The cheek-teeth are broad and heavy, with the salient angles on the inner sides of the upper molars and the outer sides of the lower molars squarely truncated. The lowest Microti retain many primitive characters not found among the highly specialized genera which alone are known to represent the Lemmi. But in one important respect the most primitive voles stand upon a higher plane than any of the lemmings. The lower incisor in all voles has pushed its way back- wards through the jaw to a point considerably behind mg, a character which indicates much later divergence from the primi- tive Murine stock by Microti than by Lemmi. The shaft of the lower incisor, lingual to the molars anteriorly (as it is through- out its course in the lemmings), crosses obliquely below or be- tween the roots of m2 and m,^ to the labial side of the jaw, where the growing base of the tooth invades and finally, in the most INTERRELATIONSHIPS OF GENERA 41 highly specialized forms, colonizes the condylar process. If we admit that the great backward extension of the incisor is a more Fig. 21. — Ondatra zihelhica Limiaeus. Skull and right mandibular ramus dissected to show the alveolar courses of the teeth; the bottom figure shows the left mandibular ramus dissected from the outer side (enlarged). ancient feature in voles than is the hypsodonty of the molars, the peculiar relation of the molars to the lower incisor, just described and so characteristic of all Microti, is easily explained ; 42 MICROTIN^ the molars as they have become hypsodont have invaded deeper levels of the jaw, and there they have had to mould themselves and adjust their curvatures to the shaft of the incisor already present. In the lemmings it is otherwise; the molars have become hypsodont much sooner, probably before the growing base of the incisor had pushed back to the level of TO2. Among the voles Evotomys is one of the most primitive genera, although its species show a fairly wide range in the degrees of specialization which they have severally obtained. In this genus, and in the closely related but more progressive genera discussed below, the skull is characterized by the structure of the bony palate, which terminates behind as a simple, thin-edged, transverse shelf, not provided with a postero-median sloping septum ; deep postero-lateral pits are, however, present, but they pass forwards freely under the edge of the shelf. The palate of Evotomys is thus very similar to that of Dicrostonyx and Lemmus, and strikingly dissimilar in appearance from that of Synaptomys and the higher voles such as Arvicola and Microtus. The auditory bullae are always well inflated, thin walled, and simple, and in nearly all species they lack all trace of internal spongy tissue. The cheek-teeth are of limited growth, develop- ing two roots each in adult stages of wear ; in pattern they are more or less simplified or reduced longitudinally; cement is present in their re-entrant folds. In outward form the species show few signs of specialization in any particular direction. In the most primitive species (members of the glareolus group) the skull is rather lightly and delicately built, with the temporal ridges weakly developed and widely separated in the interorbital region, and with the post-orbital crests of the squamosals small though distinct. The cheek-teeth are small and the incisors slender ; the lower incisor passes from the lingual to the labial side of the jaw as in all voles, but its passage between mg and m^ takes place at such a low level in relation to m^ that this tooth is not noticeably displaced. The dentinal spaces of the molars are characteristically confluent, semi-opposed and not perfectly alternating as in the higher voles, and the salient angles, in adult stages of wear, are peculiarly rounded ; m^, to^, TOj, and m^ are reduced as in normal species of Microtus ; m^ in its least-reduced form presents five salient angles on each side ; m^ is typically reduced to a pattern somewhat resembling that seen in the OTj of such voles as Microtus nivalis, and consists of a posterior loop, five triangles, and a short anterior loop. This pattern is an old one in the genus, dating from at least the Upper Pliocene, and in normal members of the glareolus group young examples of the nil show no trace of ephemeral complications in the anterior loop ; but in some far-eastern forms traces of a fourth outer fold, which is reduced by insulation of its internal part, may be seen. At the other end of the series, though apparently connected with the more primitive forms by a long series of gentle grada- INTERRELATIONSHIPS OF GENERA 43 tions, are the members of the E. rufocanus group. These are specialized for subsistence upon a coarser and tougher diet than that used by the glareolus group. The general size of the animal is increased. The cheek-teeth have become taller-crowned and more robust ; the alveolar capsules of irfi and m^ rise up promi- nently in the sphenorbital fissure, subsiding only in old age, after the failure of the dentinal pulps and the formation of roots to the teeth ; m^ is encapsulated and noticeably displaced by the shaft of the lower incisor. In pattern the teeth are somewhat modified ; the inner and outer salient angles acquire a perfect alternation and in middle stages of wear the dentinal spaces are tightly closed, although they are confluent in young stages of wear and again acquire an " Evotomys-l^ke " confiuency in extreme ojjl age ; m^ is reduced in adults and has only three salient angles on each side ; but when young both it and m-^ show commonly ephemeral complications. One character in which the members of this group are rather more primitive than are most of the other species of Evotomys may be mentioned; m^ and m^ in young stages of wear frequently show traces of cusp n. In correlation with the larger and heavier molars the temporal muscles have become stronger, and under their combined influence the skull has acquired a massiveness and angularity which resembles that seen in many species of Microtus. In the interorbital region, which becomes more constricted, the temporal ridges approach each other with age and in extreme old age may even come in contact, although usually they remain separated by a deep and narrow sulcus. AscHizoMYS, known from only one individual, is certainly a remarkable member of this group in which the outward form has become lemming-like. The skull is described as being essentially as in normal Evotomys, broad, depressed, lightly built, smooth and rounded. The cheek-teeth are said to be persistently growing, but with the general pattern and rounded salient angles so characteristic of less specialized Evotomys ; m^ is small and weak with a well-developed fourth outer angle; in the mandible njg is encapsulated and displaced by the incisor as in the E. rufo- canus group. These cranial and dental characters, coupled with my own experience of the younger stages of growth of E. rufo- canus, lead me, however, as explained below, to suspect that Aschizomys is not of generic value, but that it is based upon an adolescent specimen of a more or less aberrant member of the E. rufocanus group. EoTHENOMYS is apparently descended from some primitive form of Evotoinys. Externally the only important modification is seen in the mammary formula which is reduced from the normal 2 — 2 = 8 to 0 — 2 = 4. In all essential respects the skull resembles that of Evotomys. The cheek-teeth have become completely hypsodont and rootless, and in the lower jaw m^ is noticeably displaced by the incisor. In one respect the molars 44 MICROTIN^ are more primitive than those of Evotomys ; in most of the species cusp n is largely developed in 7n^ and nv^, in which teeth it forms a large extra postero-internal salient angle. Anteliomys is a closely related genus which goes a little further than Eothenomys in three respects : — the temporal ridges tend to fuse in the interorbital region, the palate develops a conspicuous though horizontal median spinous process poste- riorly, and cusp 11 is more reduced in vi^ and m- than is usual in Eothenomys. In one respect it is more primitive; m^ is much more complex, having five or even six salient angles on each side. The first outer fold of this tooth is usually shallow, leaving the first outer triangle confluent with the anterior loop — a character seen in other genera, such as AUicola. Alticola is also an offshoot from some primitive form^f Evotomys, as is clearly shown by the structure of the palate and the general form of the molars. The cheek-teeth have become rootless, and m^ is noticeably displaced by the incisor. The molars are characterized by the great width of the re-entrant folds, which contain very little cement, and by rather perfect alternation of the inner and outer elements ; these features give the teeth in most species a peculiar long-drawn-out appearance. In pattern m}, irfi and all the lower molars are about as reduced as in normal Evotomys; irfi shows from species to species an interesting series of gradations, beginning with forms in which the tooth is about as complex as in Anteliomys, and ending with others in which there are only three outer and two inner salient angles ; the first outer fold of this tooth is, as already mentioned, almost invariably shallow. The skull is lightly built and in essential respects closely resembles that of Evotomys. The temporal ridges are widely separated in the interorbital region; the post-orbital squamosal crests are moderately developed; in the palate the lateral bridges are frequently incomplete and the hinder edge is often furnished with a blunt median spine. The auditory bulte are simple, thin walled, and sometimes greatly inflated; they are destitute of spongy tissue within, and the stapedial artery is naked. The species are more or less highly specialized for life at high altitudes; the more primitive, with relatively complex teeth, normal Evotomys-\We essential external characters, including long, thinly-clothed tails and naked soles, occur at lower elevations; the more specialized forms, with simplified teeth, short and densely clothed tails, and hairy soles, inhabit higher regions. Some remarkable species, inhabiting the bare talus slopes of Central Asia, have acquired remarkably flattened skulls fitting them for life in rock crevices ; these have been referred to a special subgenus Platycranius by Kascenko, but apart from the peculiar flattening of the skull there is nothing to distinguish them from the more specialized forms of Alticola. Hyperacrius is apparently a peculiar descendant from some form of AUicola. The genus has become specialized for INTERRELATIONSHIPS OF GENERA 45 fossorial life; the fur is short and dense and in one species {H. wynnei) it is highly modified and mole-like ; the eyes and ears are reduced, the fore-claws are slightly lengthened, the tail is shortened, but fully clothed; the mammary formula is reduced to 1 — 2 = 6. The temporal muscles are greatly in- creased in size and strength and the reduction of the eyes is no doubt partly correlated with this muscular increase. In the skull the temporal ridges fuse, at an early age, to form a linear median interorbital crest, and the squamosals, frontals, parietals, and interparietal all show a characteristic temporal modification ; the post-orbital crests of the squamosals are greatly lengthened, extending from the interorbital region outwards and backwards on each side almost to the glenoid articulation. The anterior palatal foramina are reduced to more or less shortened, narrow slits. The palate in the smaller species is essentially as in Alticola ; but in the largest form {H. wynnei) the posterior median sloping septum is represented by a short spine, in form rather like that seen in Anteliomys, but differing in that it is not hori- zontal but sloping ; if this spine were directly connected laterally with the post-palatal pits, the palate in H. wynnei would be similar to that of many species of Microtus ; as it is the pits have already very definite, salient, inner borders, which run forwards to effect a junction with the dorsal surface of each lateral bridge near the base of the spine ; thus the structure of the palate in this species may be said clearly to foreshadow the palate of the higher voles. The presphenoid is reduced to a slender bar in all species of Hyperacrius, and a similar reduction is to be seen in some but not in all species of Alticola. The auditory bullae are as in Alticola as regards essential structure; but they are consider- ably smaller and less inflated. The cheek-teeth are light, rootless and tall-crowned ; they have normally differentiated enamel, but their re-entrant folds lack cement. The enamel pattern is essentially as in Alticola; but m^ is simplified and its posterior end is more reduced and shortened than in any species of Alticola. DoLOMYS apparently represents or at least is very nearly related to forms which must have been directly ancestral to the Water Voles (Arvicola). Its fossil remains have been found, hitherto, only in the Upper Pliocene of Hungary; but recently a living species has been discovered on the mountains of Montenegro, where, secure from competition, the genus has survived until the present day. This living species is a large, soft-furred, long- tailed vole, bearing such a close outward resemblance to the Alpine Voles {Microttis (Chionomys) nivalis) that it was at first mistaken for a member of that group. As in other primitive voles the cheek-teeth are rooted in adults ; indeed, in this genus the teeth appear to be more brachyodont than is usual in such genera, for even in the living species the cement spaces and pulp cavities close at the bases of 46 MICROTIN^ the teeth in individuals so young that they have not yet fully acquired the adult pelage. In pattern the teeth are not very different from those of Evotomys ; but mj, which possesses five alternating triangles between the posterior and anterior loops, shows in young stages of wear some vestigial and ephemeral complications of the anterior loop ; whereas m^ is reduced and has only three or four outer and three inner salient angles. Cement is present in the re-entrant folds of the teeth in the recent species, but is not developed in those of the Pliocene forms. The genus is most satisfactorily distinguished from Evotomys and placed upon a higher plane by its skull. The hinder portion of the palate shows a very broad and ill-defined median slop- ing septum, small, shallow and indefinite postero-lateral pits, and more or less incompletely developed lateral bridges. In Dolomys, therefore, we see the beginnings of that palatal struc- ture, which is better developed in Arvicola, and becomes perfected in the most specialized species of Microtus. The temporal ridges, in the recent species at all events, fuse anteriorly to form a median interorbital crest in adults and posteriorly they are fairly closely approximated. The alveolar capsules of the cheek-teeth do not rise up in the floor of the orbit or in the sphenorbital fissure. The auditory bullae lack internal spongy tissue, but they are large and considerably inflated; the stapedial artery is enclosed in a bony tube as far as the stapes. The mandible is normal; m^ is not noticeably displaced by the shaft of the incisor. The upper incisors are strongly curved and sometimes show a slight trace of an anterior groove. Apistomys, described from the Upper Pliocene of Hungary and known only from fragmentary remains, is scarcely to be distinguished as a genus from Dolomys. In it m^ has a general " arvaloid " appearance, whereas in Dolomys this tooth is more like that of some forms of Evotomys or Microtus nivalis in general form. MiMOMYS is a very interesting and important genus, repre- sented by numerous species in the Upper Pliocene of Europe and by one species in the early Pleistocene of Britain. It appears to have descended from some primitive species of Dolomys, i.e., from a form more primitive in skull structure than is the only living representative of the latter genus. So far as is known, the temporal ridges remained rather widely separated in the inter- orbital region in Mimomys, instead of fusing into a weak linear crest as in recent Dolomys. But in other respects Mimomys has gone much further than Dolomys, and in the later Pliocene and earlier Pleistocene deposits of Britain species of Mimomys shade off imperceptibly into forms which cannot be distinguished with available materials from the genus Arvicola. The palate is essentially as in Arvicola, its postero-median sloping septum being short, broad, but rather well defined. As in Dolomys the cheek-teeth of Mimomys develop roots in INTERRELATIONSHIPS OF GENERA 47 adult stages of wear. But as we trace the genus onwards from earlier to later horizons we observe that the species become more and more hypsodont, the growth of the molars ceasing and roots being developed at later moments in the life of the individual in the more modern species than in the earlier members of the genus. Simultaneously the teeth undergo a process of simplifi- cation, visible not only as we trace the genus forward from one geological horizon to another, but as we follow the development or wear of the individual tooth from infancy to age. The genus thus presents us with some beautiful examples of recapitulation which are described in detail below (p. 111). The most interest- ing tooth, as always in voles, is wij. In the most primitive forms it no doubt resembled the corresponding tooth of Dolomys in possessing not fewer than five closed triangles between the posterior and anterior loops ; in addition the third outer salient angle was complicated by a large vestige of one of the tubercles of the primitive median row giving rise to a peculiar feature which I have called the " prism-fold." In early species, like M. plioccenicus, the third outer re-entrant fold becomes reduced in early middle age by the conversion of its internal portion into an enamel islet, and this islet and the " prism-fold " persist until a very advanced stage of wear has been reached. When very young the m^ of a late species, e.g., M. intermedins, resembles a young m-^ of M. plioccenicics in all essential respects. In exactly the same way it is reduced by the conversion of the third outer fold into an enamel islet, and as wear proceeds the islet disappears and a simplified pattern showing only three triangles between the posterior and anterior loops is brought to light on the surface of the crown. This simplified pattern is the characteristic adult pattern of the tooth in the later species, and it is developed and perfected long before the tooth ceases to grow. The archaic elements in the later species are confined to the apical portions of the crown, and the process of reduction, though in every detail a faithful repetition of that seen in M. flioccenicus, is performed within the first few weeks or days of the individual's existence instead of occupying most of its life as in the older form. In adult stages of wear the cheek-teeth of the later species of Mimomys are not to be distinguished from those of Arvicola by pattern alone, and as the general progress is towards hypsodonty and persistent growth there comes a moment in the geological scale when, with existing materials, it is impossible to say definitely whether we are dealing with Mimomys or with Arvicola. If the teeth could be shown to develop roots in old age they would be referred to the former genus; but as no sign of roots, not even an incipient closure of the cement spaces or pulp cavities, can be detected, the remains in question, from the later Cromerian beds of Bacton and the early Middle Terrace deposits of the Thames, have to be referred to Arvicola. 48 MICROTIN^ Arvicola is thus shown to be a direct descendant of Mimomys ; but the species now so widely spread over Europe and Asia are probably not to be regarded as the immediate offspring of M. intermedius and allied British species ; for certain dental reasons it is more probable that they have come down from close relations of M. intermedius which once existed in Eastern Europe or Asia. The living members of the genus all show well-marked fossorial specialization which tends to become extreme in members of the A . scherman group ; in the more familiar riparian species slight specializations fitting these voles for aquatic habits have been superimposed upon fossorial characters. Modern s'peciesoi Arvicola are large voles with massive, strongly ridged, and angular skulls when adult. The temporal ridges fuse in the interorbital region, and the squamosals, with well- developed post-orbital crests, approach each other, encroaching upon the frontals anteriorly as age advances The palate posteriorly is now essentially as in normal voles; the pterygoid fossae are deep. The auditory bullae are rather small but now show a slight development of spongy tissue within, and the stapedial artery is completely enclosed in a bony tube. In the more fossorial species the occiput is characteristically inclined, pressed forwards above, and the upper incisors are noticeably straightened and protruding, the animal evidently using the skull and incisor teeth as one of its chief tools in digging. The cheek-teeth are persistently growing, with normally differentiated enamel, and with cement present in the re-entrant folds. The enamel pattern is characterized by the simplicity of m^, with only three salient angles on each side, and of m^ in which there are only three closed triangles between the posterior and anterior loops. In very young stages of wear the anterior loop of m^ shows some extremely interesting ephemeral complications which are described at p. 107. In this genus, as in many other members of the Microtinae, persistent growth of the cheek-teeth appears to be accompanied by persistent growth of the skeleton; in the oldest indi- viduals examined, among the enormous amount of fossil and recent material at my disposal, not only are the molars still in vigorous growth but the epiphyses of the limb-bones are still unfused with their shafts (Plates III and IV). Apparently, that is so far as actual observation goes, voles of this genus are animals that never stop growing and never grow old. But no doubt, if one could keep the vole alive in natural conditions, but secure from the fatal stroke of accident, a time would come when cheek- teeth and skeleton would cease to grow and the animal would become senile and die in the normal manner. Phaiomys is an interesting genus restricted to the highlands of Central Asia, in which region it is usually found inhabiting the damp meadows bordering the Alpine water-courses. It is closely related to Arvicola, and seems like the latter to be descended INTERRELATIONSHIPS OF GENERA 49 Fig. 22. — Phaiomys leucurus Blyth. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size. 50 MICROTIN^ from some close ally of Mimomys intermedius. The cheek-teeth have become persistently growing, and m^ is a little reduced ; but otherwise their pattern and particularly that of the mj, is exactly as in adult M. intermedius. In the skull the temporal ridges fuse to form a median interorbital crest, but the squa- mosals, which have small but rather conspicuous post-orbital crests, remain widely separated anteriorly. The palate is essen- tially as in normal Microtus, but its postero-median sloping septum is rather short and broad. Traces of the median suture of the palate and of the sagittal suture of the interparietal some- times persist. The auditory bullee are well inflated and their walls are strengthened by a thick development of dense spongy tissue ; the mastoid portion is also slightly inflated, and the stapedial artery passes through the stapes enclosed in a bony tube. Externally Phaiomys is evidently modified for fossorial Fig. 23. — Phaiomys leucurus Blyth. Crown views of cheek-teeth : a. right upper, 6. right lower molars. habits and for life in its special station; the fur is peculiarly long and soft; the tail is shortened and densely clothed; the palms and soles are overgrown with hair, and the claws, both of the hands and feet, have been considerably lengthened. The North American genus Phenacomys appears to repre- sent, in its essential characters, the common stock from which two of the most important groups, namely, Pitymys and its associates and Microtus and its closest allies, have descended. In external appearance most of the living species of Phenacomys do not differ much from typical voles; they possess moderately large eyes, well-developed ears, normal hands and feet, and a moderately long, well-clothed tail. Two or three forms, some- times placed in a special subgenus Arborimus, have, however, remarkably long tails, and they retain the arboreal habits which may have characterized the long-tailed ancestor of all voles. The skull resembles that of the more primitive species of Microtus in general form, having a moderately long rostrum, INTERRELATIONSHIPS OF GENERA 51 rather broad interorbital region, and oblong and depressed brain- case. The temporal ridges are widely separated in the inter- orbital region and behind; the squamosals have moderately salient post-orbital crests and are widely separated anteriorly, although in old age they show a slight tendency to encroach upon the frontals and approach each other. The upper border of each jugal bone is convex, giving a fusiform expansion to the central part of the zygomatic arch. The upper portion of the infraorbital canal is more spacious than usual. The palate posteriorly is formed nearly as in Microtus, but the postero- lateral bridges are usually absent, the postero-median septum is short and horizontal, and the postero-lateral pits are very shallow ; these features of low relief in the palate may be correlated with the brachyodonty of the cheek-teeth, and cause the palate of adult Phenacomys to resemble that of Microtus and other more highly developed voles in young stages of growth. The pterygoid fossise are shallow, their floors being nearly flush with the ventral surface of the basisphenoid. The palate, choanse, basisphenoid, and basioccipital are broad. The auditory bullae are small, globular, simple and without internal spongy tissue ; the stapedial artery is naked as it approaches the stapes. The mandible is nearly normal, but m^ is not displaced by the shaft of the lower incisor, which passes below m^ to terminate in the base and at the hinder margin of the condylar process below the dental foramen; the groove between the alveoli of the cheek-teeth and the ascending ramus is not " pocketed " posteriorly by the alveolar sheath of the lower incisor as in Microtus, but remains open behind as in Evotomys. Although some primitive characters are thus to be observed in the skull and mandible of Phenacomys the claim of the genus to a lowly position among voles rests chiefly upon the evidence of the cheek-teeth. These are very light and of limited growth, each molar developing two roots when adult. The enamel shows the beginning of a normal differentiation, being slightly thicker on the concave than on the convex sides of the salient angles. The re-entrant folds are narrow and transversely deep, giving to the teeth a characteristic longitudinally crowded appearance, but they do not contain cement. The enamel pattern is peculiar, characterized especially by the approximate equality of the inner and outer salient angles and infolds in upper molars and by their disparity in lower molars, in which the inner salient angles are conspicuously larger and the inner infolds much deeper than those of the outer side; m^, m^, wig and ?»3 are essentially as in normal voles, the last-named tooth being, however, in living species of Phenacomys, usually a little more reduced than usual, its antero-external angle becoming obsolete ; w^ is also noticeably reduced in living species, having only three salient angles on each side; ?% is a complex tooth with four or five outer and six inner salient angles, consisting 52 MICROTINiE of a posterior loop, followed by from three to seven sub- stantially closed alternating triangles, and terminated by an anterior loop of crescentic shape formed chiefly by the Fig. 24. — Pitymys planiceps Miller. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size (B.M., No. 10.7.7.219, type and only known specimen). fifth and sixth inner salient angles ; in young stages of wear the outer side of the anterior loop shows some minute and quite ephemeral complications. The interest of this tooth in Phena- comys lies in the unusual variability which it displays, from I INTERRELATIONSHIPS OF GENERA 53 species to species, in the number of closed triangles ; in some species only the three posterior are closed, those in front being Fig. 25. — Neodon sikimensis Hodgson. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size. more or less confluent with each other and with the anterior loop, the tooth presenting an appearance which elsewhere is characteristic of genera like Pitymys ; in other species four, five, six, or even seven triangles may be closed, the tooth resembling 54 MICEOTIN^ that of genera like Microtus. Among other voles {Mimomys, Arvicola, Pitymys, Microtus, etc.), with the exception of Orthri- omys and Herpetomys, the possession by m^ of three closed triangles only, on the one hand, or of four or more on the other, has been a distinction of generic importance since Pliocene times at least; but in Phenacomys both types of m^ are associated, and the distinction between them, if it be of any systematic value at all, is of no more than specific importance. Ancient species of Phenacomys, in which the outward characters may have been a little more primitive, those parts of the skull under the direct influence of the temporal muscles a little less modified, and a little less reduced than they are in the living species of this genus, may thus well have been the ancestors common to both Pitymys and Microtus, with their respective adherents; indeed, such forms may well have been ancestral to all the Fig. 26. — Cheek-teeth of Pitymys subterraneus de Selys-Longchamps. Crown views : a. right upper, b. left lower molars. other known voles as well. Unfortunately, nothing is at present known of the palaeontological history of Phenacomys. Pitymys is an ancient genus, now widely distributed in Central and Southern Europe and in South-eastern North America. The earliest remains are apparently those obtained from the later Pliocene (Cromerian Beds) of Britain; these consist of teeth and fragmentary lower jaws, indicating the presence of several species, which in size and form agree in all essential respects with the teeth and lower jaws of recent Pitymys. But whether these ancient fossils belonged to Pitymys, as we now understand it, or to Neodon, Tyrrhenicola or some other genus, cannot be determined with the material at present available. The living members of the genus are small voles with very small eyes, small ears, short tails, and soft, short, dense and more or less mole-like fur. They have large hands provided with rather long claws, relatively short feet with claws rather shorter than those of the fingers, moderately hairy soles and INTERRELATIONSHIPS OF GENERA 55 only five plantar tubercles. Flank glands are present in adult males. In the typical subgenus the mammary formula is reduced to 0 — 2 = 4 ; in the subgenus Micrurus, which is slightly less modified, the mammary formula is 1 — 2 == 6. Fig. 27. — Cheek-teeth of Pitymys majori Thomas. Crown views : a. right upper, h. left lower molars. The general external characters have thus been considerably modified in relation to the subterranean habits of these voles. The skull is more or less evidently modified for fossorial life, usually with a rather smooth, delicately built, more or less depressed braincase. The temporal ridges are weakly developed Fig. 28. — Cheek-teeth of Pitymys ibericus Gerbe. Crowa views : a. right upper, b. left lower molars. and are widely separated in the iuterorbital region ; the post- orbital crests of the squamosals are usually feeble. The palate is normal and the pterygoid fossae are deep. The auditory bullae are usually swollen, their mastoid portions also being inflated; the walls of the buUos are strengthened by spongy tissue and the stapedial artery is enclosed in a bony tube which passes through the stapes. The mandible and incisors are normal. The cheek-teeth are rootless, with normally differentiated 56 MICROTIN^ enamel, and cement is present in the re-entrant folds. The enamel pattern is closely similar to that of Microtus, but m^ has only three closed triangles in front of the posterior loop, the fourth and fifth triangles being more or less broadly con- fluent with each other and with the anterior loop. In some species w^ shows some interesting phases in its reduction. Fig. 4a Fig. 5a Fia. 29a.— Cheek-teeth of Neodon. a. Right upper molars. 1. N. sikimensis Hodgson (B.M., No. 15.9.1.218). 2. N. forresti Hinton (type). 3. N. irene Thomas (B.M., No. 12.3.18.13). 4. N. carruthersi Thomas (B.M., No. 9.4.3.93). 5. N. oniscus Thomas (B.M., No. 11.11.1.3). In face of the severe competition offered to it in Europe and N. America by the very numerous species of Microtus, Pitymys has been forced to adopt fossorial habits; it owes not only its continued existence, but its outward and cranial specialization, and probably its present wealth in species, entirely to these habits. In those parts of the highlands of South-eastern Central INTERRELATIONSHIPS OF GENERA 57 Asia, where no species of Microtus occurs, members of the Pitymys group have been able to persist on the surface of the ground, leading the lives of normal Microtines, and they have undergone a process of cranial specialization exactly parallel with that which, in other places, has been undergone by the species of Microtus. These peculiar Asiatic representatives of the genus Pitymys constitute the genus Neodon. Hg.lb. Fig. 5b Fig. 296.— Cheek-teeth of Neodon. b. Left lower molars (same specimens as in Fig. 29a). In Neodon the fur is soft and full, but is not highly modified as is usual in Pitymys; the moderately large ears, provided with a distinct antitragus, are evident above the fur ; the tail is moderately long and is fairly well clothed ; the fore and hind claws are about equal ; usually the sole has six pads, although these are sometimes reduced to five. The mammary formula is the primitive one, 2 — 2=8. The skull is nearly as in Microtus. The temporal ridges fuse in adults to form a weak but linear median interorbital crest; the squamosals, frontals, and parietals are correspond- ingly modified with age. The palate is normal. The auditory 58 MICROTIN^ bullae have only a weak development of spongy bone within. The dentition is essentially as in Pitymys, m^ difiering from that of Microtus in having only three closed triangles in front of the posterior loop. Tyrrhenicola, a fossil genus known from the Pleistocene deposits of some of the Mediterranean islands, is very closely related to Neodon. Only one well-marked species is known, a large vole with a skull in which the facial portion is unusually long and narrow, and the braincase is lofty and subcylindrical. The temporal muscles were evidently powerfully developed, the temporal ridges fusing in the much constricted interorbital region to form a median linear crest; the post-orbital processes of the squamosals are moderately developed. The pterygoid fossae are very deep, indicating powerful pterygoid muscles. The palate is highly specialized, resembling that of " Stenocranius " in the Fig. 30. — Cheek-teeth of Pedomys haydenii Baird. Crown views : a. right upper, 6. left lower molars. genus Microtus. The auditory bullae are small. The incisors are rather slender, but are normal in other respects. The cheek- teeth are persistently growing, and in other respects are essentially like those of Pitymys and Neodon. Pedomys is a North American genus apparently closely related to Neodon. It is slightly more highly specialized than the latter externally, having long, coarse, not specially modified fur, small ears, concealed in the fur, but provided with a large antitragus, broad hands and feet, with the hind-claws slightly longer than those of the hand, with five plantar tubercles, and with hairy soles; the tail is short; the mammary formula is reduced to 1 — 2 =^ 6. The skull and teeth are essentially as in Neodon, but the mastoid portions of the rather small auditory buUaj are considerably inflated. Orthriomys and Herpetomys are two very closely related genera, each represented by a single species of restricted dis- tribution, the one stranded upon Mount Zempoaltepee, Mexico, INTERRELATIONSHIPS OF GENERA 59 at altitudes above 8000 feet, the other upon the mountains of Guatemala at altitudes above 9000 feet. They have apparently descended from a primitive Phenacomys-like stock and have Fig. 31. — Cheek-teeth of Orthriomys urribrosus Merriam. Crown views : a. left upper, 6. left lower molars. evolved in the direction of true Microtus; but although they have gone too far in certain respects to be regarded as members of this genus, or as representing its ancestors, the char- acters suggest that all three may have been derived from a common Phe7iacomys-\ike ancestor. The cheek-teeth in both Fig. 32. — Cheek-teeth of Herpetomi/s guatemalensis Merriam. Crown views : a. left upper, b. left lower molars. Orthriomys and Herpelo»iys have become completely hypsodont and persistently growing; the temporal muscles have become stronger than in the parent form, producing in the adult skull more salient ridges which fuse anteriorly to form a median interorbital crest. In the general character of the enamel pattern of its molars, small size of the auditory bullae, general 60 MICROTIN^ structure of the skull, and longer tail, Orthriomys remains nearer to Phenacomys than does Herpetomys, which in the features mentioned makes a nearer approach to Microtus. In both genera the number of closed triangles in front of the posterior loop in m^ varies between three and five, and in this they agree with Phenacomys, in which m^ displays a similar though wider (3-7) variability; elsewhere in the group we find that the posses- sion of only three closed triangles by m^ is a most constant generic character of ancient standing dating, in the case of Pitymys, at least from the Pliocene period, and that variability in the number of closed triangles in m^ is shown only by those genera in which less than four never occur. In the atrophy or complete suppression of the inguinal mammae, Orthriomys and Herpetomys differ rather strikingly from other members of the subfamily, in which usually the inguinal pairs are constant and the pectoral pairs are liable to reduction. The flattening of the skull in Tig. 33. — Cheek-teeth of Proedromys Thomas. Crown views : a. left upper, 6. right lower molars. Orthriomys is, no doubt, as in the genus Pitymys, a specialization correlated with its more fossorial habits. Of the genera constituting the Microtus-gxovi^, Proedromys, known from a single specimen collected in Western China, is in two respects the most primitive. Its upper incisors are very broad, recurved, and grooved, and its lower incisors are remark- ably short, scarcely invading the condylar process behind. But in the other characters of the skull and teeth it is quite highly speciaUzed. The cheek-teeth are rootless, tall-crowned, and broad, with normally differentiated enamel, and with cement in the re-entrant folds. The enamel pattern is generally as in Microtus, but m^, m^, and 3H3 are somewhat more reduced than usual; m^ has only three outer and two inner salient angles, its posterior loop formed mostly by the third outer angle; m^ has only four closed triangles, the fifth triangle or fourth inner angle being broadly confluent with the short, rounded anterior loop; in m^ the third outer angle is obsolete. The skull is massively built. The temporal ridges probably fuse in extreme INTERRELATIONSHIPS OF GENERA 61 Fig. 34. — Microlus calamorum Thomas. Dorsal, ventral, and lateral views of subadult skull, enlarged ; the small figure shows the skull in dorsal view, natural size. 62 MICROTIN^ old age in the interorbital region; the post-orbital processes of the squamosals are prominent and peg-like, but the squamosals Fig. 35. — Microtus hartingi Barrett -Hamilton. Dorsal, ventral, and lateral views of subadult skull, enlarged ; the small figure shows the skull in dorsal view, natural size. do not appear to approach each other or to encroach upon the frontals with advancing age. The palate is normal; the INTERRELATIONSHIPS OF GENERA 63 auditory bullae contain spongy tissue. Externally Proedromys is a very ordinary looking vole with long coarse fur, a moderately Fig. 36. — Microtus ratticeps K. & Blasius. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size. long tail, and six plantar tubercles. The mammary formvda is 2 — 2 = 8. Microtus is the largest and the most widely distributed 64 MICROTIN^ genus in the subfamily, comprising a very large number of recent and fossil forms and ranging over most of the Palsearctic and Fig. 37. — Microtus orcadensis sandayensis Miller. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size. Nearctic regions. The oldest known remains are those found in the later Pliocene (Cromerian Beds) of Britain; but these INTERRELATIONSHIPS OF GENERA 65 consist chiefly of more or less fragmentary jaws and isolated teeth, representing three or four species, and do not throw any Microtu8 (Slenocranius) ravidulus Miller. Dorsal, ventral, and lateral views of skuU, enlarged ; the small figure shows the skull in dorsal view, natural size. particular light upon the ancestry or relationships of the genus. It is more convenient to reserve the discussion of the various subgeneric groups into which the numerous species of Microtus V.L. F 66 MICEOTIN^ fall for detailed treatment under the genus. None of the species is specially modified for fossorial or aquatic habits. All have normal coats, moderately developed eyes and ears, and normal hands and feet, with claws of normal length, those of the hind-foot being the longer, with moderately hairy soles and with six plantar Fig. 39. — Cheek-teeth of Microtus roherii Thomas. Crown views : a. right upper, 6. left lower molars. tubercles ; the tail is short, or of medium length, and the mammary formula is 2 — 2 = 8 (except in the small mexicanus group, where it is reduced to 1 — 1 = 4). The skull is normal, varying in shape, massiveness, and angularity with the species. The temporal ridges, stronger or weaker, fuse (at all events in old age) in the interorbital region; the post-orbital processes of the Fig. 40. — Cheek-teeth of Microtus calamorum Thomas. Crown views : a. right upper, h. left lower molars. squamosals are more or less developed and an anterior encroach- ment of the squamosals upon the frontals is more or less evident. The palate is of normal type in all, and in some forms becomes highly specialized by the extension of the post-palatal pits and the attenuation of the postero-median septum. The pterygoid fossae are deep. The auditory bullae contain spongy tissue and the stapedial artery is enclosed in a bony tube. The mandible INTERRELATIONSHIPS OF GENERA 67 and incisor teeth are normal. The cheek-teeth are endowed with persistent growth, and have the enamel normally and sometimes conspicuously differentiated, thick enamel forming the concave sides of the salient angles, thin enamel forming their convex sides ; cement is present in the infolds. In pattern m^, tn^, mg, and m^, are normal, with or without clear traces of cusp n; m^ has from three to five salient angles on each side; mj consists of a posterior loop, usually followed by five substantially closed triangles and terminated by an anterior loop ; in some forms the fifth triangle (fourth inner angle) is confluent with the anterior looj) ; in others, in which the anterior loop is extraordinarily complex, additional triangles may be closed off from the base of the loop so that the number of the closed triangles rises to six, seven or even eight. Chilotus, represented by a few species in North America and Asia, is very closely related to Microtus, but it is somewhat Fig. 41. — Cheek-teeth of Microtus clarJcei Hinton. Crown views : a. right upper, b. left lower molars. modified for more fossorial habits, or j^ossibly for creeping under logs or stones. Its skull is curiously depressed, and the mastoid portions of the auditory buUiB are distinctly inflated. The fur is short and dense, without any admixture of stiff hairs. The ears are rather small, and the plantar tubercles are reduced to five. In all other respects it agrees with typical Microtus. Lasiopodomys is a small but rather remarkable Asiatic genus. Closely related to Microtus, as is shown by the almost typical skull and teeth, it is considerably specialized externally for fossorial habits, somewhat in the manner of Lemmus or Phaiomys. The fur has become soft and fine ; the ears are short, scarcely appearing above the fur, and are nearly naked. The fore- claws are considerably lengthened, and the thumb is armed with a sharp claw instead of the usual flattened nail. The soles of the hind-feet are densely haired, and although six pads are present, the two posterior are very small, placed low down, and completely hidden beneath the hair. The tail is short and densely clothed. The skull is broad and rather flat, with normal 68 MICROTIN^ palate; the temporal ridges fuse anteriorly in old age. The auditory bullae are densely spongy within, and their mastoid portions are noticeably inflated. The persistently growing cheek-teeth are substantially as in Microius; but m^ has only three salient angles on each side, and m.^ has the third outer angle obsolete; ?% is of characteristic form, with a posterior loop, five closed triangles, and a small anterior loop of peculiar squarish shape. Lagurus, widely distributed in Eastern Europe, Asia, and Western North America, is a remarkably isolated, in some respects a very primitive, in others a highly specialized genus. It appears on the whole to be most closely related to Microtus and its nearer allies, and no doubt traces its descent from some primitive Phenacomys -like vole. The external form is highly modified for fossorial habits, the general appearance of the animal being very Lemming-like. The fur is long and very soft. The eyes are moderate. The ears, without an antitragus, are very small and hidden in the fur. The hands and feet are short and broad ; the thumb bears a small pointed nail ; the claws are of moderate length, the hind ones very slightly the longer; there are five plantar tubercles, but they are completely con- cealed beneath the dense hairy covering of the soles. The tail is very short and is densely clothed, the hair forming a short terminal pencil. The mammary formula is 2^2 = 8. The short and broad skull has a remarkable superficial resemblance to that of Dicrostonyx in dorsal view. The rostrum is moderately short and broad, and the zygomata leave its sides squarely, without any prezygomatic notch. The interorbital, region is moderately constricted, but the temporal ridges, although very salient in adults, are rather widely separated anteriorly by a deep median sulcus much as in Dicrostonyx; the ridges are rather closely approximated behind, compressing the interparietal, which is abruptly truncated laterally. The squamosals are widely separated anteriorly, with rather small but prominent peg-like post-orbital processes. The palate is essentially as in Microtus, with a long postero-median sloping septum. The pterygoid fossse are short and deep, their floors lying at a level considerably dorsal to the ventral surface of the basi- sphenoid. The auditory bullse are very large and are highly modified, with the external meatus shortly tubular and much straitened; the cavity is partly filled with a dense sponge of bone ; the mastoid portion and tegrnen tympani are similarly enlarged and spongy, the mastoid part bulging outwards rather conspicuously between the paroccipital process and the lateral process of the supraoccipital ; the stapedial artery is enclosed in a bony tube. The mandible is normal, although the angular process is rather small ; the lower incisor displaces m.^ as usual. The upper incisors are strongly curved or slightly " opisthodont." The cheek-teeth are persistently growing and exceedingly INTERRELATIONSHIPS OF GENERA 69 tall-crowned, their alveolar capsules rising high in the floors of the orbit and sphenorbital fissure. The enamel is differentiated Fig. 42. — Lagurus luteus Eversmann. Dorsal, ventral, and lateral views of skull, enlarged; the small figure shows the skull in dorsal view, natural size (B.M., No. 12.4.1.121; the median dorsal suture is quite obliterated in thia specimen). as in Microtus, but the re-entrant folds lack cement. The enamel pattern is characterized by the great width of the re- entrant folds, perfect alternation of the salient angles, long- drawn-out appearance of the teeth, and usually by the retention 70 MICROTIN^ in m^ and m- of distinct vestiges of one or more of the " inter- mediate " tubercles ; m^ and m^ apart from the generic peculiarities are normal, but the second inner fold of m^ and the inner fold of m- are usually complicated by a vestige of the " protoconule " (cusp y) ; m^ is more or less simplified, with three or four outer and two or three inner salient angles. In general appearance the upper molars, particularly m^, show a strong likeness to those of Alticola. In the mandible w^ has five closed triangles between the posterior and anterior loops as in Microtus ; m^ and m^ are normal, but the last-named tooth has its outer angles and infolds less reduced than usual. Notwithstanding its high specialization for aquatic life Ondatra, peculiar to North America, is apparently more closely related to Phenacomys than to any other Microtine genus, and it has descended from some primitive Phenacomys-like stock. In one respect, the palate, it is even a little more primitive than Fig. 43. — Cheek-teeth of Lagurus luteus Eversmann. Crown views : a. left upper, 6. right lower molars. is any known Phenacomys. The cheek-teeth are distinctly primitive and Phenacomys-like ; they develop roots when adult and have enamel of about equal thickness on the concave and convex sides of the salient angles ; the re-entrant folds are partly filled with cement and close the dentinal spaces rather tightly ; the depth and narrowness of these folds, coupled with the great breadth of the crowns, impart to the teeth an appearance of longitudinal crowding. The enamel pattern of m^, m^, and m^, is normal ; m^ has three or four salient angles on each side ; m^ is a complex tooth with a posterior loop, five to seven closed triangles, and a short anterior loop which shows remarkable ephemeral complications when unworn (see p. 115). In m^ the outer salient angles are somewhat reduced in size, but the outer infolds are deep, so that the first pair of triangles following the posterior loop are substantially closed, and the second pair are not more widely confluent than are the corresponding triangles of W2- Obscure traces of " intermediate " tubercles are frequently to be seen in adult or well-worn teeth. INTERRELATIONSHIPS OF GENERA 71 In the skull there are two primitive features; the palate is abnormal posteriorly, lacking postero-lateral bridges, having the postero-median septum represented merely by a slender and free spinous process, and having the postero-lateral pits small and ill defined ; the auditory buUaj are small and lack internal spongy tissue, although the external meatus is distinctly tubular, and the stapedial artery is enclosed in a bony tube as far as the stapes. The mandible is nearly normal, but Fig. 44. — Ondatra zibethica Linnseus. Dorsal view of skull (enlarged). it shows one primitive character ; the lower incisor passes back- wards below the molar roots to terminate in the lower part of the condylar process, displacing m^ lingually to a very slight extent. The upper incisors are normal, without grooves, and terminate in the maxilla immediately in front of the anterior root of m^ as usual (Fig. 21). In all other respects Ondatra is greatly modified for its peculiar habits. Its size has greatly increased beyond that of its terrestrial relatives. The fur is waterproof, made up of very dense, long, fine, and silky under fur, overlaid and more or less completely 72 MICROTIN^ concealed by the longer, stiffer, and peculiarly glossy contour hairs. The eyes are small. The ears, provided with a distinct antitragus, are small, densely haired, and almost completely con- cealed in the fur. The hands are nearly normal and moderately large, with naked palms ; the fingers, including the small thumb, all bear long claws. The feet are relatively large, and are slightly twisted, metatarsal IV forming most of the anterior border of the foot (in its natural position) ; as is common in Fig. 45. — Ondatra zibeiJiica Linnseus. Ventral view of skull (enlarged). aquatic mammals digit IV tends to be slightly longer than digit III, which is ordinarily the longest toe; all the toes bear claws which are stouter and sometimes longer than those of the fingers ; the soles are naked, with five or four plantar tubercles, of which the postero-internal is very large and elongate; the normal postero-esternal pad is not developed, and the pad which in other voles is usually present between the bases of digits III and IV is sometimes present, sometimes absent. The lateral borders of the foot and of each toe are furnished with conspicuous " swimming-fringes " of stiff hairs, INTERRELATIONSHIPS OF GENERA 73 Fig. 40. — Ondatra zibeihica Linnaeus. Lateral view of skuU, enlarged ; the small figure shows the skull in dorsal view, natural size 74 MICROTIN^ and similar though weaker fringes are developed along tlie margins of the hands. The tail is very long, about two-thirds of the head and body length, laterally compressed, covered with small scales, and most inconspicuously clothed over its general surface with short stiff hairs ; but along the mid-dorsal and mid-ventral borders the hair is longer and denser, com- pletely hiding the skin, and producing behind a very short stiff terminal pencil. The mammary formula is 1 — 2 = 6. Perineal glands, secreting a powerful musk, are well developed. Apart from the primitive characters above mentioned the skull is highly specialized; it is very large and massive, and in general form and structure closely resembles the skull of the more highly specialized species of Microtus. The interorbital region is greatly constricted. The temporal ridges in adults are relatively closely approximated throughout, fusing anteriorly to produce a sharply salient, median interorbital crest. The squamosals are large, relatively closely approximated both anteriorly and posteriorly, with their post-orbital crests strongly developed and extensive, forming the square shoulders of the braincase. The rostrum is long and slender, to accommodate the large incisors. The pterygoid fossae are deep, their floors dis- tinctly dorsal to the ventral surface of the basisphenoid. Neofiber is another North American genus of great interest. It is represented by a single species known to occur only in eastern and central Florida. Like Ondatra it is specialized, though less profoundly, for aquatic life; but in its dental characters it has progressed much further than that genus. Neofiber is a very large vole, although considerably smaller than Ondatra. The fur, eyes, and ears are substantially as in the latter, but the long contour hairs produce a sort of dorsal keel in the neighbour- hood of the rump. The hands and feet are moderately large, with naked palms and soles ; the hind claws are distinctly longer and stouter than those of the fingers; the foot is less evidently twisted than in Ondatra, and there are five plantar tubercles, the postero-external one being suppressed and the postero-internal pad rounded. The " swimming -fringes " on the borders of the feet and toes are much weaker and less conspicuous than in Ondatra, but that along the outer border of the hand is better developed than in that genus. The tail is long, as in Ondatra, but is of quite different form, being terete and fully clothed with long stiff hairs. The mammary formula is 1 — 2 = 6. Flank glands are conspicuously developed in both sexes, even in the young. The skull is essentially like that of Ondatra, but is character- ized by its great fronto-palatal depth, evenly convex antero- posterior dorsal contour, and conspicuously salient post-orbital squamosal processes. The temporal ridges are more widely separated, the intertemporal portions of the parietal and inter- parietal being relatively broad. The palate is essentially as in INTERRELATIONSHIPS OF GENERA 75 Ondatra, but the postero-lateral bridges are more developed, though slender and incomplete. The choanse, pterygoid fossse and auditory bullse are as in Ondatra. The mandible and incisors are normal ; the shaft of the lower incisor displaces Wg lingually. The cheek-teeth have acquired the power of growing persistently, although in other respects (enamel, tight closure of dentinal spaces, presence of cement in infolds, and general pattern) they closely resemble those of Fig. 47. — Neofiber alleni True. Dorsal view of skull (enlarged). Ondatra. The enamel pattern of m^, ?rtj, and ni^ is, however, somewhat simpHfied ; 7n^ has only three salient angles on each side ; ?»j has a posterior loop, five closed triangles, and a short anterior loop (which shows more or less ephemeral and vestigial traces of reduced and obsolete elements) ; in m^ the antero- external angle is suppressed, although the triangles following the posterior loop are substantially closed. Two genera of extraordinary interest, namely, Protnetheomys and EUobins, remain for discussion. Certain characters, partly primitive, partly progressive, shared in common, show that these two genera are closely related ; but in spite of this close 76 MICROTIN^ relationship and of the fact that each has been profoundly modified for fossorial habits, they are very unlike. Each has solved the problems of fossorial life in its own peculiar way. Among the characters common to the two genera may be mentioned the brachyodonty of the cheek-teeth, which develop roots in adult stages of growth ; the peculiar reduction of m| ; the simplification of ?«,j ; the absence of cement from the re- entrant folds of the teeth ; the remarkable relation of the roots Fig. 48. — Neofiber alleni True. Ventral view of skull (enlarged). of mg to the shaft of the lower incisor. The lower incisor passes from the lingual to the labial side of the jaw between OTj and m^ as in other voles ; but the implanted part of m^ curves sharply backwards and follows the curve of the dorsum of the incisor upon its lingual aspect, instead of continuing straight down transversely across the shaft of the incisor. Prometheomys, represented by a single species living in the mountains of the Caucasus, is in many ways the more primitive genus. It is a rather large vole, with long and soft fur, very small eyes, and moderately large, naked and rounded ears which are provided with a large antitragus and are not concealed by the fur. INTERRELATIONSHIPS OF GENERA 77 The hands bear enormous claws, those of the three central digits being especially lengthened, slender and recurved; the thumb Fig. 49. — Neofiher alleni True. Lateral view of skull; the small figure shows the skull in dorsal view, natural size. bears a short, stout, curved claw ; the palm is naked between the pads, of which there are the usual five. The foot is also provided with large claws, but these are only half as long as those of the 78 MICROTIN^ hand; there are five plantar tubercles; the sole is naked between the pads, but is densely haired posteriorly. The tail is about half the length of the head and body, thick, and densely clothed with hair. The mammary formula is 2 — 2 = 8. In general form the skull is much as in other voles, with moderately long and broad rostrum, rather sharply constricted interorbital region, squarish braincase, normal zygomata, and with the infraorbital canal and its outer wall of the usual form. The rostrum is not unusually shallow; the occiput is vertically truncated. One primitive feature of the dorsal surface is worthy of special remark ; traces of the sagittal suture can be seen, even in old adults, between the nasals, frontals and parietals, and posteriorly the suture usually persists, dividing the remarkably small interparietal into two halves. In adults the temporal ridges are fused to form a sagittal crest which ultimately extends back- wards from the middle of the interorbital region to the occiput ; this crest reaches its greatest height on the posterior part of the frontals, and in all but the very oldest specimens the crest subsides in the mid-parietal region and the ridges, thence feebly marked, diverge slowly to the occiput, crossing the interparietal obliquely ; the inter- parietal, at all times small in Prometheomys, diminishes with age. The squamosals are very large, forming the whole dorso-lateral surface of the braincase, their upper borders being parallel; the lateral wings of the parietals, so constantly present in voles, are obsolete; the post-orbital crests, though fairly extensive, are very weak, and the squamosals anteriorly show no special tendency to approach each other by encroaching upon the frontals. These characters are correlated with the great development of the temporal muscles, of which the middle and hinder portions are especially strong ; a parallel development is seen in Ellobius, and a somewhat similar, but, for Microtinse, far less abnormal condition occurs in Ondatra. The alveolar capsules of the cheek- teeth are, in young specimens, protuberant in the floors of the orbit and the sphenorbital fissure, but these protuberances subside in adults as the crown stumps are pushed out of the capsules by the more slender roots; the capsule of m^ is nearly clear of the braincase. The antero-palatal foramina are moderately large. The palate is abnormal behind, more primitive than in any other member of the subfamily, but foreshadowing the palate of Ellobius, etc. ; it is sculptured in low relief, with broad and complete postero-lateral bridges, but with the median septum represented only by a broad irregular nasal spine ; the floors of the homologues of the postero-lateral pits lie slightly dorsal to the general surface of the palate, but they are cut off from it by the lateral extensions of the fore-part of the mesopterygoid fossa ; each pit is formed as usual by the ventral surface of the palatine bone, but each is perforated by a large foramen which occupies fully half the area of the pit. By filling up the foramina, deepen- ing the pits, placing their edges in continuation with the nasal INTERRELATIONSHIPS OF GENERA 79 Fio. 50. — Prometheomys schaposchnikowi Satunin. Dorsal, ventral, and lateral views of skull, enlarged ; the small figure shows the skull in dorsal view, natural size. (In this specimen the median suture of the interparietal has been obliterated.) 80 MICROTIN^ spine, and tunnelling under the postero-lateral bridges we would convert the palate of Prometheomys into one like that of Microtus. Fig. 5\.— Prometheomys schafoschnihowi Satunin. Skull and left mandibular ramus dissected to show the alveolar courses of the teeth (enlarged). The pterygoid fossae are fairly large but are not deep, their floors being about level with the ventral surface of the basisphenoid ; INTERRELATIONSHIPS OF GENERA 81 anteriorly they are perforated and placed in rather broad connec- tion with the orbit — a very unusual feature in voles. The auditory bulla} are moderately large, with short tubular external meatus ; the legmen tympani articulates with the squamosal in front and the mastoid portion is not inflated ; the cavity of the bulla does not contain spongy tissue, but its walls are strengthened by several incomplete perpendicular septa; the stapedial artery is not enclosed in a bony tube. The mandible is normal; the lower incisor produces a strong alveolar hump upon the outer surface of the condylar process well below the level of the condyle. The upper incisors are strongly curved, their front surfaces flush with the nasal tips and their growing bases lying in the masillaries immediately in front of m^ ; they are broad teeth, each with a weak groove on the outer third of its anterior surface ; the enamel is stained bright yellow. The lower incisors are not peculiar. The cheek-teeth are rooted, each tooth with two roots; in m^ a vestige of a third root, supporting the second inner angle, such as occurs in some of the earlier species of Mimomys, is sometimes present ; in m\^ the two roots may coalesce. The enamel is rather thick throughout, but shows feeble traces of a normal differentiation. Cement is not present in the rather wide re-entrant folds. The outer folds both in upper and lower molars tend to leave the opposed dentinal spaces confluent with each other. The enamel pattern in general is like that of Ellobius ; m^, m^, and m^ are essentially normal ; m^ has three outer and two inner salient angles, divided into two lobes by the second outer and single inner infolds ; its outer salient angles are very much reduced, with the first outer triangle very small and broadly confluent with the anterior loop, an exaggeration of the tooth form seen in Alticola and Hyperacrius ; the opposed tooth m^ is also greatly reduced and divided into two lobes, of which the posterior consists of the posterior loop proper and the first inner triangle, which are broadly confluent in consequence of the reduction of the first inner fold; the anterior lobe is formed by the second outer and third inner salient angles, broadly confluent with each other ; the third outer angle is obsolete ; m^ has three outer and four inner salient angles, consisting of a posterior loop, three substantially closed triangles, and an anterior loop formed by the broadly confluent fourth and fifth triangles which open into the short anterior loop proper. Young specimens of the molars show traces of the " intermediate " tubercles and other ephemeral complications (see p. 117). Ellobius is widely distributed in Eastern Europe and Asia, and its range, as shown below, appears formerly to have extended into Northern Africa as far west as Tunis and Algeria — the only Microtine known from the African continent. Like Protnetheomys it is highly specialized for fossorial habits ; but its mining opera- tions are conducted in a totally different way, and because the methods employed by the two animals are and have always V.L. G 82 MICROTIN^ been different, fossorial habits moulding what is essentially one and the same primitive stock have in the end produced two totally different results. Prometheomys digs its quite shallow but extensive burrows with its hands and throws out the earth in small heaps like a mole ; its powerful incisors are used chiefly for the purpose of cutting roots which may obstruct its progress, and for cutting the roots of the subalpine grasses upon which it feeds. The hands, feet, and fur are the organs therefore chiefly affected by the fossorial habits of Prometheomys; these habits have not directly influenced either its skull or its teeth to any great extent, and the chief cranial modifications of the genus are correlated with the peculiar development of the temporal muscles and therefore with the peculiarities of the food. Ellohius, on the other hand, uses its incisor teeth and skull as a powerful shovel and drill, in the manner of Spalax; accordingly the head and fur are highly specialized in this genus, but the hands and feet remain comparatively unmodified. In Ellobius the fur is very fine, short, dense, and mole-like; the eyes are small ; the ears are reduced to a mere fold of naked integument surrounding the external meatus and completely hidden in the fur. The hands and feet are of moderate size, armed with very small claws, and with naked palms and soles ; there are five palmar and six plantar pads as usual, the plantar pads being all moderately developed or small. The lateral borders of the hands and feet are fringed with stiff hairs, those on the outer margins being especially well developed. The tail is very short, and is fully clothed with stiff hairs, which form a long thin terminal pencil. The skull shows extreme fossorial specialization, being cuneate in profile, with straightened and far-protruding incisors, extremely shallow and slender rostrum, depressed anterior nares, flattened dorsal surface, and slightly inclined occiput. The zygomata are moderately stout and diverge posteriorly. The orbito-temporal vacuities are rather small ; the interorbital region is short and wide, about as broad as the rostrum. The temporal ridges are closely approximated throughout, tending to fuse or in old age fusing to form a sagittal crest extending from the interorbital region backwards to the occiput as in Prometheomys. The interparietal is very small, its sutures obliterated in old age. The squamosals are large but do not show any tendency to encroach upon the frontals anteriorly; their post-orbital crests are distinct, but the shoulders of the braincase fall away from the interorbital region much less abruptly than in Prometheomys and normal voles. The infraorbital canal is modified; its upper portion for the transmission of the infraorbital portion of the masseter medialis muscle is somewhat enlarged, and the lower slit-like portion normally serving for the transmission of the nerves and vessels has disappeared. The rostrum in ventral view is long and slender ; the anterior palatal foramina are very small. The palate INTERRELATIONSHIPS OF GENERA 83 is essentially as in Microtus, the postero-median septum being sharply defined and the postero-lateral pits deep and extensive. Fig. 52. — Ellobius fuscocapillus Blyth. Dorsal, ventral, and lateral views of skull, enlarged; the small figure shows the skull in dorsal view, natural size. The pterygoid fossse are deep, their floors being slightly dorsal to the ventral surface of the basisphenoid, and they are closed anteriorly. The auditory bullse are very small and but slightly 84 MICROTIN^ inflated ; their external apertures are very small and their walls are strengthened by a stout and wide meshwork of bony trabe- FiG. 53. — Elldbius tancrei Blasius. Lateral and ventral views of a skull dissected to show the alveolar courses of the teeth (enlarged); a. marks the position of the growing base of the upper incisor and in the ventral view points to the maxillary fenestra. culiE ; the stapedial artery is enclosed throughout in a bony tube. The mandible has lofty recurved coronoid processes, and the angular processes are reduced in each ramus to a mere ridge INTERRELATIONSHIPS OF GENERA 85 bordering the alveolar sheath of the lower incisor as it ascends the condylar process. The incisor teeth are highly specialized; they are much lengthened and straightened, forming smaller segments of much larger circles than usual ; they are without grooves, and, as is very commonly the case in Rodent incisors showing fossorial adapta- Ellohius lancrci Blasius. Right mandibular ramus dissected to show the alveolar courses of the teeth. The small figure represents the left mandibular ramus in oblique posterior view and shows the relation of the alveolar protuberance of the incisor to the condyle (enlarged). tion, their enamel is white and they tend to be rounded in transverse section. The upper incisors extend much further back- wards in the maxillary bones than in any other Microtine genus, their alveolar capsules terminating just dorsal to the surface of the palate between the molars at a point near the hinder end of m^ ; in some species the ends of the capsules actually appear on the surface of the palate, a fenestra by absorption of bone being opened in the maxilla before the advance of the growing tooth. The lower incisors are also very long, ascending the condylar 86 MICROTIN^ process to a point just below the condyle, where their capsule projects, and rises to about the level of the condyle itself. The cheek-teeth are rooted in adults. The enamel is rather thinner than in Prometheomys and about equally thick on the concave and convex sides of the salient angles. Owing partly to the shallowness of the re-entrant folds, which do not contain cement, and partly to the imperfect alternation of the inner and outer angles, the dentinal spaces show an unusual degree of con- fluence. Apart from their generic peculiarities m}, m^, and mg are in pattern essentially as in normal voles, but m^ occasionally shows traces of the cusp x^ and of the fold which primitively separates that cusp from 6 — a very archaic character; m^ is greatly reduced in much the same way as in Prometheomys, but the reduction goes even further, the first outer triangle and the Fig. 55. — Cheek-teeth of Promelhcomys scJia poscJmikotvi Satunin. Crown views : a. left upper, b. right lower molars. infold primitively separating it from the anterior loop being reduced to microscopic vestiges, and although m^ and 7)1^ develop two roots each m^ acquires but a single fang. In the lower jaw Ml consists of a posterior loop, three to five alternating triangles, and a short anterior loop ; its posterior loop is usually substantially closed in front, but the triangles are more or less confluent with each other and with the anterior loop ; the degree to which the fourth and fifth triangles are difterentiated from the base of the anterior loop varies with the species, the individual, and with age ; ^nj is essentially as in normal voles apart from its confluence ; wij is like JWg, but is a little more reduced, its antero-external angle being obsolete ; m^ and ?».2 have two roots each, which pass down through the jaw to the labial side of the incisor; m^ has a single fang, lying on the lingual side of the incisor, but sharply curved backwards to follow the dorsal curvature of the incisor, instead of continuing nearly straight downwards as in all other voles with the exception of Prometheomys. This abrupt backward I INTERRELATIONSHIPS OF GENERA 87 curvature of the tooth and its root is an exaggeration or accentua- tion of the gentle backward curvature of the m^ seen in nearly all voles ; in those in which the teeth have acquired the power of persistent growth the curvature becomes marked and it is one of the means of keeping the teeth " keyed " together tightly at the grinding surface; but what the mechanical significance of this character is in Prometkeomys and Ellobius I am not prepared at present to say. That it is a character indicating, like others, a special relationship between the two genera, I do not doubt; it may be a character of accommodation, the m^ having come into relation with the shaft of the incisor at a different moment in its developmental history from that in which these two teeth have come into relation with each other in other voles. " Bramus barbarus " Pomel,^ from the Quaternary Phosphorites I Fig. 56. — Cheek-teeth of Ellobius talpinus Pallas. Crown views : a. left upper, b. right lower molars lettered according to the homologies of the cusps. of Tunis, is a most interesting fossil form known to science only from the original description. Miller,^ in 1896, thought it " prob- able that Bramus is the type of a group differing too widely from any of the recent Microtince to be united with them in one sub- family; " but in 1918 Miller and Gidley ^ placed the genus in the family Rhizomyidsp, forming for it a special subfamily Bramince. After carefully studying Pomel's most lucid though unfortunately unillustrated account of the fossil species, I have come to the conclusion that Bramus must be regarded as a synonym of Ellobius. Every detail mentioned by Pomel, namely, the form of the molars, the characters of their roots, the relation of m^ to the shaft of the lower incisor, the form of the mandibular angle, the character of the infraorbital foramen and the form of the interorbital region, appears to be exactly as in Ellobius. The occurrence of this 1 Pomel, Comptes Rendus, Paris, 114, p. 1159, 1892. « MuxEE, N. Amer. Fauna, No. 12, p. 74, 1896. 3 MiLLEE and Gidley, Journ. Washington Acad. Sei., 8, p. 438, 1018. 88 MICROTINiE genus in the Pleistocene or even at the present day in North Africa would, of course, be in complete harmony with what is now known about the faunistic relations of North Africa with South-eastern and Central Asia. e. Key to the Genera of Microtin^. Group Lemivii. Lower incisor short, wholly lingual to the molars, and terminating posteriorly in the horizontal ramus opposite or in front of the alveolus of vi^. I. Cheek-teeth rooted in adults. (None yet discovered.) II. Cheek-teeth rootless and persistently growing, with well- clifferentiated enamel. A. Cheek-teeth longitudinally complex; with the inner and outer salient angles and re-entrant folds approximately equal in size or depth ; re-entrant folds without cement ; ml with four sahent angles on each side; m^ with four outer and three inner saUent angles; postero- external angle (" cusp n ") in m^ and m^ much smaller than the other salient angles; m^ with at least five outer and six inner saUent angles, and with at least seven closed triangles in front of its posterior loop. Skull with prominent peg-Hke post-orbital processes ; temporal ridges separated in interorbital region by a median sulcus. Auditory bullae with dense internal sponge of bone. Ears vestigial and concealed. Hands and feet short and broad ; palms and soles densely clad with crisp hair ; palmar and plantar pads vestigial; fore-claws greatly enlarged and apparently double in winter; thumb with minute nail. Tail shorter than foot, densely clothed, with a long terminal pencil. Mammae 2—2=8 Dicrostonyx. B. Cheek-teeth longitudinally simpUfied ; inner saUent angles of upper, and outer salient angles of lower molars smaller than those of opposite sides; re-entrant folds with cement; m^ with three saUent angles on each side ; m^ with three outer and two inner saUent angles ; rn? apparently consisting of four transverse loops; wij with three outer and four inner saUent angles and with three closed triangles only in front of the posterior loop. Molars broad-crowned; some of the inner saUent angles in upper teeth and of the outer saUent angles in lower teeth presenting an appearance of square truncation. Skull with shelf -like post-orbital crests; temporal ridges fusing sooner or later in the interorbital region. Thumb with large flattened, strap-shaped nail. 1. Palatal surface of skull not terminating posteriorly as a simple transverse shelf; its median spinous process I I KEY TO GENERA gg forming an inclined septum between the postero- lateral pits (as in Microtvs). Upper incisors strongly curved, each with a well- marked anterior groove; second and third transverse loops of w3 separated chiefly by second outer infold, hkull lightly built, long and narrow. Auditory bulliB large and globular, internally with a wide meshwork of bony trabecule. General outward form vole-Hke ; ears evident above fur; palms and soles naked between pads; palmar tubercles 4, plantar tubercles 6; claws and thumb- nail not speciaUy enlarged; tail slightly longer than hind-foot, rather thinly clad Synaptomys. a. Lower molars with closed triangles, their outer infolds and sahent angles well developed. Upper incisors very broad. Rostral part of skuU very stout; palate without long posterior spine. Mammae 1-2= 6 . . . Subgenus Synaptomys. b. Lower molars without closed triangles; their outer infolds and salient angles not well developed. Upper incisors not very broad. Rostral part of skull slender; palate with lontr posterior spine. ° Mamm^2_2 = 8 . . Subgenus Mictomys. ralatal surface of skull terminating behind as a simple * transverse shelf; its median spinous process when present horizontal and ending freely (as in Evotomys). Upper incisors less strongly curved, without well- marked anterior grooves, and with much hollowed tubular worn surfaces; second and third transverse loops of m? separated by first inner and second outer folds. Skull relatively massive, broad and depressed Mammte 2—2=8. i. External form vole-Uke, but more thick-set than in Synaptomys; fore-claws and thumb-nail not specially enlarged, and skeleton of hand not specially modified. Skull rather lightly built ; zygomata less widely expanded ; squamosals relatively widely separated in front. Auditory bulla? globular . . . Myopus 1. External form highly modified for fossorial habits Ears hidden in the fur. Hands large and broad the claws and thumb-nail greatlv enlarged; meta- carpals shortened and ungual phalanges lengthened . for support of claws. Feet short and broad, with large claws. Palms and soles densely haired- palmar and plantar tubercles vestigial. Tail short, robust, densely clothed, and with a long terminal pencil. Skull massive, broad and depressed; zygomata very heavy, strong, and widely expanded ;' squamosals more closely approximated in front. Auditory bulls ovate or pyriform, not greatly inflated- densely spongy within Lemmas 90 MICROTIN^ GeOUP MlCEOTI. Lower incisor long, passing from the lingual to the labial side of the molars between the bases or roots of m^ and WI3, and ascending for a greater or less distance behind the molars to terminate within or near the condylar process. I. m§ not extremely reduced, never conspicuously smaller than mf ; implanted portion of m^ not sharply recurved and never approximately concentric with the alveolus of the lower incisor. A. Skull with palate terminating behind as a transverse shelf, with or without a median spinous process; the latter when present free, its tip never connected with the inner borders of the postero-lateral palatal pits. Auditory bullae without spongy tissue. ^ 1. External characters never modified for aquatic habits; size small or medium. Skull with the temporal ridges usually separated in the interorbital region of the adult (Hyperacrius excepted). Auditory biiUse thin- walled; stapedial artery naked as it passes through stapes. Dentinal spaces of cheek-teeth not always substan- tially closed; m^ never with closed triangles. a. Cheek-teeth rooted in adults; m^ with four outer and five inner angles. External form normal. Mammae 2—2=8 Evotomys. 1. Cheek-teeth relatively weak; dentinal spaces of jrtj and m, more or less confluent in middle Ufe; m^ not encapsulated or noticeably displaced by the shaft of the incisor. SkuU not exceptionally massive or angular. Normal Evotomys. 2. Cheek-teeth robust; dentinal spaces of ?Wi and m^ tending to be tightly closed in middle life; m^ noticeably displaced by the shaft of the incisor and encapsulated on the lingual side of the jaw. SkuU massive and angular, resembling that of Microtus in general appearance . rufocanus group.^ b. Cheek-teeth rootless and persistently growing. 1. External form Lemmuig-Uke. Cheek-teeth tall- crowned but Ught, in pattern resembling those of normal Evotomys; m^ displaced by the shaft of the incisor and encapsulated . . Aschizomys.' 2. External form essentially normal, not Lemming- hke. Lower molars nearly as in normal Evotomys. a. Skull with the temporal ridges not fused in the interorbital region; post- orbital crests of squamosals never very long; anterior palatal foramina not greatly reduced. Auditory bullae never very small. Re-entrant folds of the cheek-teeth partly filled with cement. External ^ The only exception to this is afforded by Evotomys kennardi, a fossil species from the Pleistocene of Britain. The bullae of this form contain a little spongy tissue. ^ Owing to the existence of intermediate forms it is not possible to separate the rufocanus group subgenerically from normal Evotomys. 3 As to the validity of Aschizomys see pp. 43, 279. KEY TO GENERA 91 characters not highly modified for fossorial hfe; ears moderately or well developed; plantar tubercles 6 ; fore-claws not noticeably lengthened ; fur never highly modified. a. Mammse 0—2=4. *. ml usually with four, m^ usually with three well-developed salient angles on the inner side; m^ with three or four salient angles on each side, its first outer infold usually deep. Skull relatively broad, the temporal ridges feeble and widely separated in the mterorbital region; palate without con- spicuous spinous process behind. ** 1 J o , , Eothenomys. *'. m' and m- normal, the former with three, the latter with two well-developed inner saUent angles; m^ very complex, with five or six salient angles on each side, its first outer infold usually shallow, leaving the anterior loop and first outer triancle broadly confluent. ° Skull long and narrow; temporal ridges * stronger, tending to approach in the inter- orbital region ; palate with well-developed, horizontal, median spine behind. Antelio'mvs p. Mammae 2—2 = 8. Cheek-teeth with characteristic long-drawn- out appearance; ml and 7?j2 normal ; ?«» variable, sometimes complex, sometimes simplified, with from three to six outer and from two to five inner salient angles, its first outer infold usually shallow and its posterior loop very long and narrow in the more reduced forms of the tooth. Skull with the temporal ridges weak and widely separated in the interorbital region. * oi II , , Alticola. . OkuU normal, not conspicuously depressed. ** o, „ ,, , Subgenus Alticola. **. Skull greatly depressed. ,,, „ .^, ^, , , Subgenus Platycranius. bkuil with the temporal ridges fused in the inter- orbital region ; post-orbital crests of squamosals noticeably lengthened, extending backwards towards the glenoid region; anterior palatal foramina small. Auditory bullae very small. Re-entrant folds of cheek-teeth without cement; m^ and m^ normal; m^ simpfified, with three outer and two inner salient angles. Its posterior loop short and broad, its first outer infold shallow. External characters more or less evidently modified for fossorial habits; fur short and dense, sometimes mole-like; ears small- tail short. Mamma; 1-2 = 6. . . Hyperacrius. 92 MICROTINvS; 2. External characters highly modified for aquatic habits; under fur very dense and fine, contour hairs long, stiff and glossy; palms and soles naked; hands and feet furnished with " swimming-fringes " ; claws enlarged. Mammte 1—2=6. Size very large. Skull with the temporal ridges fused in the inter- orbital region when adult. Auditory bullae small ; external meatus tubular; stapedial artery enclosed in a bony tube as far as stapes. Dentinal spaces of cheek-teeth always closed; m^ with closed triangles; re-entrant folds partly filled with cement. a. Cheek-teeth rooted in adults; m^ with three or four salient angles on each side ; mj with five outer and six inner saKent angles and from five to seven closed triangles; wij with three outer sahent angles. Skull with temporal ridges closely approximated throughout ; its dorsal contour flattened from before backwards ; post-orbital processes of squamosals forming shelf-like crests. Tail laterally compressed and scaly. Feet obviously twisted; lateral margins of toes con- spicuously fringed ; fore and hind claws subequal. Contour hairs not forming a " dorsal keel " on rump Ondatra. b. Cheek-teeth rootless and persistently growing; m^ with three sahent angles on each side ; mj with four outer and five inner salient angles and five closed triangles; JWg with two outer salient angles only. Skull with temporal ridges -widely separated on the braincase ; its dorsal contour convex from before backwards; post-orbital processes of squamosals prominent and peg-like. Tail terete, densely haired. Feet less noticeably twisted; "swimming-fringes" less developed on borders of digits; hind-claws longer and stouter than fore-claws. Contour hairs forming a " dorsal keel " in neigh- bourhood of rump Neofiber. B. Skull with palate not terminating behind as a transverse sheK; median spinous process always present, con- verted into a sloping septum between the postero- lateral palatal pits, the inner borders of these pits always continuous with the tip or sides of the median process. Auditory bullse with or without spongy tissue. 1. Cheek-teeth rooted in adults. Bony palate with postero-median sloping septum short and broad. a. Skull with the temporal ridges fused in the interorbital region when adult ; palate with postero-lateral pits shallow and ill defined, the postero-lateral bridges sometimes absent. Auditory bulla} large with thin walls; mastoid portion shghtly inflated; stapedial artery enclosed in bony tube as far as stapes. Mandible with wij not noticeably cUsplaced by shaft of incisor. KEY TO GENERA 93 Cheek-teeth with the inner and outer salient angles and infolds approximately equal; re-entrant folds with or without cement, not unusually narrow or deep; 7«3 with three or four outer and three inner salient angles; m^ with five closed triangles 1. General form of m, " nivaloid." Palate narrower. Jixternal characters normal ; fur, feet and pads not speciaUy modified; ears moderately large- tail long ■' jj° J ' 3. General form of m, " arValoid.'" Palate 'broader. ™''^' b. Skull with temporal ridges widely separaled^'in™^'' interorbital region of adult. 1. Lower incisor short, terminating posteriorly in base of condylar process, its shaft not noticeably displacing m,. Palate ^vith postero-median septum but httle inclined, the postero-lateral pits very shallow, and the postero-lateral bridges usuaUy absent. Auditory buUse small ; stapedial artery naked. ^ Cheek-teeth characterized by the small size of the outer salient angles of the lower molars; re-entrant folds narrow (giving the teeth a longitudmaUy crowded appearance), not con- taimng cement; enamel thicker on concave than on convex sides of sahent angles; m^ with three salient angles on each side; m, complex with four or five outer and six inner sahent angles, its closed triangles varying in number from three to seven. External characters normal . . . Phenacomys ~. Lower mcisor ascending the condylar process for a greater or less distance, its shaft (at all events in later species) noticeably displacing m,. Palate with postero-median septum, postero-lateral pits and bridges well defined. Cheek-teeth with the outer and inner salient angles and re-entrant folds approximately equal- re-entrant folds of normal width and depth usually, but not invariably, containing cement; enamel usually thicker on convex than on con- cave sides of salient angles; m^ with no more than three sahent angles on each side, its second mner infold sometimes persistent, sometimes subject to reduction by insulation of its internal portion; nil never with more than three closed triangles m adult stages of wear, its third outer fold sometimes persistent, but usually reduced by msulation of its internal portion. (In older species the molar roots are developed sooner- m newer species at a later moment in the life of the mdividual ; the islet representing the internal portion of the third outer fold oi m, is a long persistent feature in the earher forms, but becomes quite ephemeral in later species.) tMimomys. 94 MICROTIN^ 2. Cheek-teeth rootless and persistently growing. Bony palate normal; usually boldly sculptured in adults. Mandible with m^ noticeably displaced by shaft of incisor. Auditory bullae with the stapedial artery enclosed in a bony tube. a. Inguinal mammae always present and functional; /»! with its fourth and fifth triangles, according to the genus, either constantly closed off from, or con- stantly open to each other. 1. Enamel pattern of m^ and m^ simphfied; m? with three salient angles on each side; rftj with three outer and five inner saUent angles, its third outer fold more or less reduced, with three closed triangles only and all parts in front of the third triangle merged in the anterior loop. Skull with the temporal ridges fused in the interorbital region when adult. External form always more or less modified for fossorial or for aquatic habits. a. Size medium to large. Skull massive, strongly ridged and angular when adult ; post-orbital crests of squamosals well developed ; squamosals tending to approach each other anteriorly with advancing age. Auditory bullae small, with a shght develop- ment of spongy bone tissue within. External form modified sUghtly for aquatic habits, but in some species rather more markedly for fossorial life. Hind-claws slightly longer than fore-claws ; palms and soles naked ; hands and feet sUghtly fringed laterally. Tail about half length of head and body. Mammae 2—2=8 Arvicola. b. Size small or medium. Skull less strongly built; post-orbital crests small but salient ; squamosals widely separated anteriorly. Auditory bullae large, their walls strengthened by dense spongy bone tissue; mastoid portion shghtly inflated. External form modified for fossorial life ; fur long and soft ; palms and soles densely haired, the pads concealed; claws unusually long. Tail short, about one-fourth length of head and body. Mammae 3(2) - 2 = 10(8). Phaiomys. ~. Enamel pattern of mj less reduced, its third outer infold always well developed, and one or more of the dentinal spaces in front of the third closed triangle always separated from the anterior loop. a. nil with only three closed triangles, the fourth and fifth triangles confluent with each other and substantially closed off from the parts anterior to them. Re-entrant folds of cheek- teeth partly filled with cement, a. Cranial and external characters modified for fossorial life. KEY TO GENERA 95 Skull with the temporal ridges widely separated in the interorbital region; post- orbital crests of squamosals weak ; braincase smooth, depressed, and delicately built. Auditory bullae well inflated, with spongy tissue within. Fur soft, dense, and more or less mole-like. Eyes and ears small, the latter concealed. Soles moderately hairy ; plantar tubercles 5 ; claws shghtly lengthened. Tail short. Pitymys. *. Mammae 0—2 = 4. . . Subgenus Pitymys. **. Mammae 1 — 2=6. . Subgenus Micrurus. p. Cranial and external characters not specially modified for fossorial life. Skull with the temporal ridges fused in the interorbital region when adult; post- orbital crests moderate; squamosals tend- ing to approach each other anteriorly with advancing age. *. Skull with facial portion not unusually elongate; palate normal. Auditory bullae strengthened by net of bony trabeculae within, f. Braincase not specially deep and sub- cyHndrical. Auditory bullae moderately large; mastoid portion not noticeably inflated. Fur soft and full. Ears evident above the fur. Claws of hands and feet rather long, subequal ; soles not densely hairy; plantar pads usually 6, but sometimes reduced to 5. Tail rather long, from one-third to half the length of the head and body. Mammae 2—2=8 Neodon. ■ff. Braincase deep and subcyHndrical. Audi- tory bullae small; mastoid portion inflated. Fur long and coarse. Ears small, con- cealed in fur. Hands and feet broad; hind-claws slightly longer than fore- claws ; soles hairy ; plantar tubercles 5. Tail shorter, rather less than one-third the length of the head and body. Mammaj 1 — 2=6 . . . . Pedomys. **. Skull with facial portion unusually elongate; palate with deep and extensive postero- lateral pits and very long and narrow postero-median septum. Auditory bullae small t Tyrrhenicola. , nil usually with five or more closed triangles, never less than four ; the fourth triangle never confluent with the fifth. o. Vestiges of " intermediate tubercles " (" pro- toconule," " metaconule," etc.) normally 96 MICEOTIN^ absent from cheek-teeth ; cement present in re-entrant folds; m^ without closed triangles. *. External characters normal, not obviously modified for fossorial habits. f. Upper incisors grooved, broad and re- curved; lower incisors short, scarcely invading the condylar process; m^ with three outer and two inner sahent angles; ^n^ with four closed triangles, the fifth triangle confluent with the short anterior loop; m^ with its third outer angle obsolete. Skull massive ; temporal ridges fused in the interorbital region; post-orbital processes prominent and peg-like; palate normal. Auditory bullae with spongy tissue within. Fur long and coarse ; tail moderately long; plantar tubercles 6. Mammae 2—2=8 Proedromys. tf. Upper incisors normally without grooves, not unusually broad or recurved ; lower incisors long, ascending the condylar process to a greater or less extent. Skull with temporal ridges fused in the interorbital region when adult; palate normal. Auditory buUae with spongy tissue within Microtus. **. External form and cranial characters modi- fied for fossorial life. Skull broad and flat; temporal ridges fusing in inter- orbital region in old age. Auditory bullae densely spongy within; mastoid portion inflated. Ears small, f. Fur short and dense; claws not length- ened; plantar tubercles 5. Mammae 2-2=8 ChUotus. tf. Fur soft and fine ; fore-claws considerably lengthened; soles densely haired; plantar tubercles 6, very small, crowded together, and concealed; tail short; m^ with three saUent angles on each side; m^ with five closed triangles and small squarish anterior loop. Lasiopodomys. p. Vestiges of " intermediate tubercles " (" pro- toconule," " metaconule," etc.) normally present in upper molars. Re-entrant folds wide, without cement; mj vidth five closed triangles; m^ with closed triangles. SkuU short and broad; temporal ridges salient but separated in the interorbital region by a median sulcus ; squamosals widely separated anteriorly, their post- orbital processes peg-like and prominent; KEY TO GENERA 97 palate normal. Auditory bullae very large; mastoid portion and legmen tympani swollen ; densely spongy within. Upper incisors strongly curved. External form Lemming-like; fur long and soft; ears very small, concealed in fur; claws moderate, the hind ones slightly the longer ; soles densely haired ; plantar tuber- cles 5, concealed ; tail very short. Mammae 2—2=8 Lagurus. *. External form more Lemming-like; tail shorter; m^ with four closed triangles and three outer saUent angles. Subgenus Lagurcs: **. External form less Lemming-like; tail longer; m^ with three closed triangles and two outer salient angles. Subgenus Lemmiscus. b. Inguinal mam'rase absent or f unctionless ; rwj with its fourth and fifth triangles indifferently closed or open, the number of closed triangles varying individually between three and five. Re-entrant folds of cheek-teeth deep and narrow, partly filled with cement; m^ with closed triangles, its first outer fold deep. Skull with temporal ridges fused in the inter- orbital region when adult. Auditory bullae with a small amount of spongy tissue witliin ; stapedial artery enclosed in bony tube. Fur long and soft ; ears large ; plantar tubercles 5. 1. Skull long, narrow, and somewhat depressed. Auditory bullae small. Cheek-teeth, particularly the upper molars, like those of Phenacomys in general appearance ; tn^ with three saUent angles on each side; mj with three or four outer and four inner salient angles; m^ with three salient angles on each side. Tail long, half length of head and body. Mammae 2—0=4. Orthriomys. 2. Skull less depressed. Auditory bullae large. Cheek-teeth of more normal general appearance; m^ with three outer and four inner salient angles ; wij with four outer and five inner salient angles ; m^ with two outer and three inner salient angles. Tail shorter, about one-third length of head and body. Mammaj 2—1 = 6 . . . Herpetomys. n. m§ extremely reduced ; m* always and m^ sometimes con- spicuously smaller than m% ; implanted portion of m^ sharply recurved, lying upon and tending to be con- centric with the lingual aspect of the dorsal surface of the alveolus of the lower incisor. Cheek-teeth rooted in adults ; re-entrant folds wide and shallow, without cement ; inner and outer dentinal spaces mostly confluent, forming or tending to form transverse pairs; enamel thick not clearly diUerentiated ; enamel pattern simplified, that of m^, vi-, and m^ essentially normal. V.L. " H 98 MICROTIN^ Skull with temporal ridges closely approximated or fusing throughout; squamosals very large, their upper borders tending to be parallel, the lateral wings of the parietals obsolete. Highly modified for fossorial life. A. Skull of nearly normal general form; interorbital region much constricted ; post-orbital crests of squamosals shelf-like though weak; infraorbital canal normal; anterior palatine foramina large; palate abnormal, inner borders of postero-Iateral pits not directly con- nected with palatal shelf, each pit with a large foramen. Auditory bullae moderately large, without spongy tissue, but their walls strengthened by incomplete perpendicular septa ; external meatus tubular ; stapedial artery naked. Mandible normal, \vith well-developed angular and obUquely inchned coronoid processes; alveolar pro- tuberance of incisor well below the level of the condyle. Upper incisors strongly curved, ending nearly flush with nasals in front and terminating in maxilla just in front of m^ behind, broad, each with a weak anterior groove ; enamel of incisors yellow, m^ bilobular, vrith three very small outer and two large inner salient angles, its first infold very shallow leaving the first outer triangle broadly confluent with the anterior loop ; mj with three outer and four inner salient angles, and three alternating substantially closed triangles; m^ bilobular with two outer and three inner saUent angles, the first inner infold shallow leaving the posterior loop and first triangle broadly confluent, the third outer saUent angle obsolete. Fur long and soft; eyes small; ears moderately large, not concealed; hands with enormous claws, those of the three central digits especially lengthened, slender, and recurved; feet with large claws but these only about half as long as those of fingers ; palms and soles naked between the pads ; plantar tubercles 5 ; tail short and thick, densely haired. Mammae 2 — 2 = 8. Prometheomys. B. Skull of extreme " fossorial " form, with straightened and far-protruding incisors, shallow and slender rostrum and shghtly inclined occiput ; interorbital region broad, not sharply constricted off from braincase ; post-orbital crests of squamosals weak; infraorbital canal wider above than usual, but lacking the lower fissure-Hke portion seen in Prometheomys and normal voles; anterior palatine foramina very small ; palate essen- tially as in Microtus. Auditory bullae small, their walls strengthened by numerous stout perpendicular trabeculse; external meatus very smaU; stapedial artery enclosed in a bony tube. Mandible with nearly vertical coronoid process and much reduced angular process; alveolar protuberance of incisor rising to the level of the condyle. Incisors with white enamel ; upper incisors unusually long, slender, and straight (" proodont "), their tips protruding far in front of the nasals, their alveolar EVOLUTION OF INCISORS 99 portions extending backwards between the molars to terminate near the palatal surface of the hinder part of the maxilla opposite the hinder end of m'. Cheek-teeth essentially as in Prometheomys, biit m' with its first outer infold still more reduced ; m^ with the first inner infold less reduced ; m^ in some species less reduced, the fourth and fifth triangles in such forms being more clearly separated from the anterior loop. Fur fine, short, dense, and mole-like ; eyes small ; ears vestigial and concealed; hands and feet of moderate size; claws very small ; palms and soles naked ; plantar pads 6. Tail very short, with a long terminal pencil. Mammae 2 — 2 = 8 Ellobius. /. Special Notes. 1. The Evolution of the Incisor Teeth. Gnawing, as is well known, is the most characteristic of all rodent habits. It is the habit which from the first has exerted the greatest and most continuous influence upon the structure of these mammals. By degrees it has produced all those essential modifications of the incisor teeth, of the jaws, of the jaw muscles, and of the bones to which those muscles are attached, which now distinguish the order Rodentia so sharply from all others. But, if this be so, gnawing must be regarded as the primitive rodent habit, and departures from that habit, whether great or small, are specializations, progressive or retrograde accord- ing to the point of view. In turn this gives rise to the reflection that incisor teeth, that by their form and curvature are perfectly adapted for gnawing are, among living rodents, more primitive than those which have lost something of this form and curvature in becoming adapted to other uses.^ In rodents with typical gnawing propensities the exact form and curvature of the incisors, no doubt, depend very largely upon the nature of the substance to be gnawn ; the hard shells of nuts, the trunks of trees, and the stems of grasses make very different demands upon the teeth. But in all such typical rodents the upper incisors are strongly curved, " orthodont " ^ at least, and * Lest this view appear far-fetched I will add that no one looking at the essential characters of a rodent incisor, namely, its life-long growth and the restricted distribution of its thick enamel, can doubt that it is primarily the product of the gnawing habit. All rodents, whatever their present habits, have descended from ancestors that gnawed vigorously. A good parallel to some of the retrogressive specializations observable among Rodentia is afforded among Camivora by Proteles; no one can fail to recognize the stamp of its predatory ancestors in its incisors and canines ; but no one would attribute predatory habits to Proteles itself after inspection of its vestigial cheek-teeth. « Thomas, Ann. Mag. N.H., [9], 1, p. 35, 1918 ; ibi?i, is large and persists as an essential element of the crown. In the other teeth this cusp has been suppressed ; but small complications of the anterior loop in m^ and m^ and of the posterior loops of the lower molars may be seen, at all events occasionally, in all genera. These features when present always occur, as will be seen from the figures, in definite situations, and on the view of the homologies here adopted they are easily interpreted as being vestiges of the cusps x^ and x ; but if any other view of the homologies be taken they are meaningless. In the unworn m^ of Dicrostonyx cusp x^ is separated from cusp 1 by a little V-shaped median and longitudinal valley which furrows the anterior surface of the tooth (Fig. 57). More or less clear traces of the separation of those two cusps may be seen in many genera from time to time. Of the median tubercles persistent traces of y and z, representing respectively the " protoconule " and " metaconule " of Trituberculy, frequently occur in the upper molars of Microtinse ; such a vestige of y is one of the normal features in the upper molars of the genus Lagurus (Fig. 43). In ?rt^ we find a striking parallel to m^. The posterior part of the tooth, corresponding with the anterior part of wij (and no doubt owing to the fact that it, like the latter, develops freely, unhampered by the presence of any neighbour), shows a great range in complexity. In many genera as in Anteliomys (Fig. 88) some species of Evotoinys (Fig. 78) and AUicola (Fig. 91, la), it is very complex ; but from that most complicated condition a continuous gradation of forms leading down to the most reduced and simplified condition, found in Prometheomys and Elldbius, may be traced (Figs. 55, 56, 91, i-i6, 95). In some genera as in Arvicola, Evotomys and some species of Microtus (e.g., 31. nivalis) where the adult 7>i^ is more or less extensively simplified longitudinally, young or unworn specimens of this tooth show more or less ephemeral additional salient angles on one or both sides behind. In species in which the cheek-teeth develop roots in old age the changes seen in the enamel pattern as the teeth are worn down are most instructive. In Evotomys riifocanus (Figs. 10, i, 2, ii ; 81, 82) «ii and m^ are characterized by their greater complexity in youthful stages of wear ; by middle life the patterns of these teeth are much simplified and the dentinal spaces in all the molars are tightly closed, the teeth having a form recalling Microtus rather than Evotomys; but after the roots are developed, in extreme old age, the re-entrant folds become shallow, the salient angles rounded, and the teeth acquire a form which resembles that seen in any ordinary and more primitive species of Evotomys. Here we have an illustration of the truth contained in that second paragraph quoted from Ameghino on p. 104. EVOLUTION OF MOLARS 119 At this point, before proceeding to a more general comparison, it will be useful to sum up the results obtained from a study of the molars of the Microtina; alone. Firstly, there is evidence that the cheek-teeth above and below have been evolved from teeth in which there were three longitu- dinal rows of tubercles. Of these rows the median has retained, in its modernized form, its full functional importance ; but the outer row in the lower and the inner row in the upper molars have suffered reduction. Secondly, there is evidence that all the molars have to a greater or less extent suffered reduction in a longitudinal direction. This reduction has affected the ends of the teeth ; it has been most far-reaching in m'f and )ii^; the anterior ends of m\ and the posterior end of m^, free from contact with neighbouring teeth, have suffered least. Apart from cusp x^, normally lost by all the teeth except m^, reduction has chiefly affected the posterior ends of the upper and the anterior ends of the lower molars ; in other words simplification has proceeded from behind forwards in upper molars, from before backwards in lower molars. The parts, which accord- ing to Fleischmann's theory are homologous in the two series are equally liable to reduction above and below ; whereas, as is familiar to all upholders of the tritubercular theory, what are generally supposed to be the homologous cusps have very different fortunes in the two jaws. Lastly there is no evidence at all of any progressive complica- tion of the molars within the group. On these various grounds we may reject the theories of the progressive complication of the teeth of Microtinse advanced or supported by many authors.^ As the result of my work I conceive the molars of the ancestor of the Microtinaj to have been low-crowned, rooted, multitu- berculate teeth, each tooth, above and below, having three longi- tudinal rows of tubercles. In this primitive form ?/i^, m| and m^ possessed at least twelve, m^ not fewer than fifteen, and m^ from eighteen to twenty-one tubercles. We now proceed to a comparison of the cheek-teeth of Micro- tinse with those of other Muridse. The oldest Muridse at present known are the Cricetodonts of early and middle Tertiary age. In the most primitive of these (i.e., those in which the cheek-teeth retain more of the primitive complexity), such as Cricelodon cardurcense Schlosser,^ described from the Miocene of France, the outer sides of the upper molars and the inner sides of the lower molars are nearly as complex as in any Microtinse, showing repre- sentatives of the cusps 1, 4, 2 (never present in Microtinte), 5, and 3 with x^ in m^. On the other hand, the median tubercles are much reduced as in modern Cricetinae, and the inner cusps of the upper, 1 TuLLBERG, Ueber das System der Nagetiere, 1899, pp. 229, 232, 235, 23'7. 239, 442-443. Miller and Gidley, Joum. Washington Acad. Sci., 8, pp. 436-^37, 1918. ^ ScHLOSSER, Die Nagcr des europaischen Tertiars, Palaeontographica, 31, p. 90, Taf. viii, figs. 28 and 35, 1884. 120 MICROTIN^ and corresponding outer cusps of the lower molars are reduced to two (6 and 7). Cusps 4, 5, 6, and 7 have become dominant and the teeth make a close approach to that form, with two large cusps (6 and 7) on one side and the remains of five cusps separated by three enamel islands or folds on the other, which is seen in so many rodents in all the three great tribes of Simplicidentata. Cricetodon has thus specialized in a direction quite different from that followed by the Microtinae ; although some of its dental characters point back to the same common source, it cannot be considered as in any way ancestral to the group with which we are dealing. The Nesomyinse now restricted to Madagascar are of especial interest, inasmuch as one of the genera, Brachytarsomys, has developed in such a way that Forsyth Major was able to describe it as " a forerunner of the Microtinae." ^ The most primitive genus is Nesomys. Its cheek-teeth are brachyodont, rooted, and complex. Like those of Cricetodon, they show cusps 1, 4, 2, 5, and 3 well developed, with a trace of x^ in in}, and cusps 6 and 7 are the chief elements on the opposite sides of the teeth. But all the cusps are about equally developed and, what is more important, the median row is well represented, not specially reduced as in Cricetodon. In this one respect Nesomys makes a nearer approach to the voles. The vole-like Brachy- tarsomys is generally regarded as a fossorial modification of the Nesomyinae in which the cheek-teeth (Fig. 63) have acquired a strikingly Microtine general appearance, although they are far more brachyodont than in any known vole. The molars agree further with those of Microtinae in lacking any trace of an independent cusp 2. The skull (Fig. 18) under the influence of fossorial habits and of jaw muscles, which have developed exactly as in some of the higher voles, has become almost the counterpart of that of ArvicoJa or Microtus in the advanced position of the orbit, the structure of the infraorbital canal, the zygomatic arch, the form and course of the temporal ridges (which fuse in front to form an interorbital crest), and the flattening of the braincase. In other respects, correlated chiefly with the extreme brachyo- donty, the skull of Brachytarsomys is very primitive. Although in both respects modified in uiuch the same manner as in the higher voles, the cheek-teeth are already too reduced and the skull is too highly specialized for the genus to be considered as representing the ancestor common to all Microtinae ; but that it has descended from that ancestor there can be little doubt and it may perhaps be necessary later on to transfer Brachytarsomys from the Nesomyinae to the Microtinae. Pending the examination of spirit material and unworn teeth it is not advisable to make such a change. Probably on account of the close superficial resemblance which exists between the adult cheek-teeth of Microtinae and those of 1 Forsyth Major, Proc. Zool. Soc, 1897, p. 719. EVOLUTION OF MOLARS 121 such highly specialized Cricetines as Neotoma, it is very generally, but in my opinion quite erroneously, believed that the Microtinse are more closely related to the CricetinjB than they are to the Murinse. Miller and Gidley ^ indeed form a family Cricetidse Fig. 63. — Cheek-teeth of Brachytarsomys albicauda Gunther. Crown views enlarged : a, h. right upper, a', h'. left lower molars arranged and lettered to show the homologies of the cusps, a, a', younger and less worn ; h, h' . older and more worn. (containing the Cricetinae (including the Nesomyinse), Gerbillinae, Microtina?, and Lophiomyinae) which they distinguish from the MuridiE (containing the Dendromyinse, Murinse, Phloeomyinae, Otomyinse, and Hydromyinse) by the structure of the upper molars. According to them the tubercles of these teeth in Cricetidse * Miller and Gidley, Journ. Washington Acad. Sci., 8, pp. 436, 439, 1918. 122 MICROTIN^ " always present a longitudinally biserial arrangement and never develop a functional third series on lingual side of crown " ; whereas in Muridse, as they define the family, the upper teeth have " a functional row of tubercles on lingual side of crown internal to the protocone and hypocone, these tubercles entering conspicuously into the plan of modification of the crowns." This classification is based, in my opinion, on a misconception ; all three rows of tubercles are j^resent in Cricetinse, Nesomyinaj, and Microtinse as well as in the " Muridse " of Miller and Gidley. The inner series (cusps 6 and 7, etc.) are present in all ; it is the median row and not the inner row that fluctuates in development. The median tubercles, as we have seen, are well developed in primitive Nesomyinae ; and in their transformed, modernized state they contribute largely to the crown of the Microtine molar. But in Cricetinse, as in the still older Cricetodonts, the median tubercles sufEer reduction and eventually disappear. In the beautiful figures of South American forms, published long ago by Hensel,^ clear traces of these median tubercles are visible (cf. Hensel's figs. 23a and b, 25a and b, and 26a and b) ; and in unworn or but very slightly worn upper molars of Neotoma small vestiges of these median tubercles occur. In Murinse the median tubercles are excessively developed, more or less at the expense of their neighbours, as was recognized by Osborn.^ The cusps which Miller and Gidley refer to as " protocone " and " hypocone," in the passage quoted, are in reality the " protoconule " and " metaconule," cusps y and z of the notation used by me. Clear trace of such an excessive develop- ment of the median row, transformed as it is, is met with in many Microtines, as for example, in the unworn m^ of Ondatra described above. Among Murinse forms like Apodemus epimelas and Chiruromys {Pogononiys) present the most primitive, least- reduced molars met with in the subfamily ; but in all, the longi- tudinal simplification of the molar crowns has proceeded further than in the lowest Microtinse. The enamel in the unworn teeth of some of the more highly specialized Murinse, as was first pointed out by Hensel,^ does not extend over the summits of the tubercles. In Microtinse a similar condition is frequently seen in unworn teeth. Atrophy of apical enamel is clearly a specialization effected in order to render the teeth serviceable for the particular task before them from the first moment that they cut the gum, a very important adaptation in such precocious animals as young mice. Summing up, there is in my opinion ample evidence to prove that all the Muridse (and indeed all the Simplicidentate Rodents) 1 Hensel, Abhandl. Konigl. Akad. Wiss., Berlin, 1872, Taf. i-iii. ' OsBORN, " The Rise of the Mammalia in North America," Address Boston, 1893, p. 19. ^ Hensel, Zeitschr. deutsch. geol. Ges., 8, p. 283, 1856. ScHLOSSER, Die Nager des europaischen Tertiars, p. Ill, 1884. EVOLUTION OF MOLARS 123 have descended from ancestors with brachyodont multituberculate molars, in which the tubercles both in upper and lower molars were triserially arranged. In the most primitive Nesomyinae and Murinae, the transverse complication, occasioned by the triserial arrangement of the tubercles, is more completely preserved than in the other subfamilies ; in the most primitive Microtinae, on the other hand, the no less archaic longitudinal complexity of the teeth is preserved to an unusually large extent, although in all other respects the teeth of some of these primitive Microtinae have reached the highest possible level of specialization. This association of archaic and progressive features in one and the same form, often in one and the same organ, is a familiar one to the student of mammals ; over and over again we find that what is funda- mentally a lowly type has preserved itself and prospered in competition with more highly organized forms, by adapting all its plastic characters to the needs of some special environment. But if the views expressed above be sound, then those brilliant generalizations constituting the famous Tritubercular Theory, which has stimulated so much productive palaeontological research, do not apply to the Simplicidentate Rodents. Equally inapplicable is the still more brilliant though less known theory of Winge. In the teeth of these mammals, whatever may be the case elsewhere, we have nowhere to deal with increasing complica- tion of the molar crown ; on the contrary in every family molar evolution resolves itself into a tale of reduction. There is through- out this great group a general tendency for the molars to become hypsodont, and increasing hypsodonty implies, as a rule, pro- gressive simplification. Forsyth Major taught me many years ago two things : — (1) that a triangular tooth is not necessarily tritubercular; (2) that we are not entitled to neglect any element of a molar, no matter how small and inconspicuous it may be, if we can identify it when it occurs. He was the first ^ to show that the primitive rodent molar must have been a very complex thing, and the first to suggest a Multituberculate origin for the whole group. The views expressed above are merely extensions of his theory; but they have been arrived at quite independently, as the result of many years' work, upon the basis of material far richer than that at the disposal of my late friend. In conclusion it may be stated, in order to put my meaning beyond doubt, that like Forsyth Major I regard the Allotheria of Marsh (= the Multituberculata of Cope) as the probable source of the Simplicidentate Rodents, as well as of all the other Placentals. Although as regards general characters members ^ The chief papers in which Forsj-th Blajor expressed and developed his views are: 1873, " Nageriiberreste aus Bohnerzen Siiddeutschlands und der Schweiz," PaU-eontographica, 22; 1893, " On some Miocene Squirrels," P.Z.S., 1893, p. 179; 1897, -'On the Malagasy Rodent Genus Brathy- vromyst" P.Z.S., 1897, p. 695; 1899, " On Fossil and Recent Lagomorpha," Trans. Linn. Soc. London, 2nd Ser. Zool., 7, p. 433. 124 MICROTIN^ of some other Orders, e.g., the Insectivora, are no doubt more primitive than Rodentia, the latter are in fundamental molar structure among the most primitive of living mammals. In no other group do we find such clear and such generally distributed traces of the original Multituberculate tooth structure. The molar types regarded as primitive by the supporters of the Tritubercular Theory, by Winge, and by the supporters of many other rival theories, are in my opinion secondary and not primitive types, although often very ancient ; these secondary forms have been developed again and again, from Jurassic times onwards, in all Orders as the result of progressive simplification of the primitive Multituberculate tooth. Although probably no known Multituberculate can be claimed as ancestral to Rodents, it is interesting to note that some of them, like Polymastodon, developed along lines which are parallel to the direction followed by the Rodentia. It is among Allotheria alone that we find molar teeth which fit the conception of the primitive Simplicidentate molar given above. The earliest Rodents known, those from the Basal Eocene, are already quite highly specialized in many ways, and the Order is evidently one of very great antiquity. 3. The Dental Formula of the Muridce. The question as to whether the anterior cheek-tooth above and below in the Muridse is a true molar (my), a premolar (j5j), or a persistent milk-molar (tojo|) has attracted attention from time to time. Owen ^ stated that in rodents " the first or anterior of the molar series, whether the number be 2 — 2, 3 — 3, or 4 — 4, is a premolar ; it has displaced a deciduous predecessor in a vertical direction." But such a replacement is entirely unknown in Muridss, and Owen's formula, pj wirl, lacking the support of any positive evidence, has been abandoned for many years in favour of that now in common use, namely, mj-:o-;f!. In 1872 Forsyth Major ^ was led to suspect that the anterior cheek-tooth, above and below, in Muridse is neither a " molar " nor a " premolar " in the ordinary sense of those terms, but is the posterior milk-molar, which in this family has become persistent in each jaw, the permanent premolar (p4), normally replacing this tooth, having been suppressed. Long afterwards, but independently when working at Microtinse, I arrived at a similar conclusion,^ and adopting Hensel's tooth notation (which I have always preferred) both Forsyth Major and I wrote the formula 1 Owen, Anatomy of Vertebrates, 1868, 3, p. 300. 2 Forsyth Major, Atti Soc. Ital. Sci. Nat., 15, p. 112 (1872); Palseonto- graphica, 22, p. 75 (1873); Abhandl. Schweiz. palaontol. Ges., 4, p. Ill, footnote 3 (1877). 3 HiNTON, Proc. Geol. Assoc, 21, p. 490 (1910). BRITISH FOSSIL SPECIES 125 dm\- m|f|, which iu the more usual notation would be expressed as mpl, mY?2- Kellogg ^ used the formula p} nij^. In his paper he cites the views of Forsyth Major and myself, but does not indicate clearly whether he intends to agree with us or not, although he seems to admit that the homologization of the front lower cheek-tooth in Microtinye with mp^ would afford an explanation of its complexity. The reasons for the suggested homologization of the teeth in question with vip^ have been given fairly fully by me.^ Briefly they are that the lower tooth is so complex in Microtinse that we are unable to derive it from any known rodent molar without postulating the addition of new parts ; that there is no evidence of increasing complication at all in this tooth ; and that every- where, even when most complex, the tooth is undergoing reduc- tion whether we trace it forwards from youth to age in one species or forwards from one species to another in successive geological horizons. Posterior milk-molars are commonly more complex than true molars ; therefore the tooth in question may be a persistent milk-molar. The facts (1) that in Mus musculus the last molars are sometimes suppressed (in South America *), indicating that numerical reduction, when it occurs, takes place from behind ; and (2) that in those rare cases in which an extra tooth is developed in Muridse it appears at the posterior end of the series (Microtus ag rest is as recorded by Wiuge * ; Saccostomus hildcB, B.M.^), so that according to the current notation the extra tooth would have to be regarded as mi ; are cited in support of this view. If, however, the view now maintained that the molars of all Simplicidentata were primitively as complex at least as they are in the most primitive Microtina3 be correct, the argument above, in so far as it is based upon the complexity of the anterior lower tooth, loses its force. In any case I can only reiterate the opinion expressed iu 1923 that no change in the commonly accepted notation of the cheek-teeth of Muridaj should as yet be adopted. 4. The Range in Time of the British Fossil Microlince. Fossil remains of Microtina3 occur abundantly in the later Pliocene and Pleistocene deposits of Britain. A study of the remains collected from the deposits of the Norfolk coast and from the terrace-deposits of the Thames shows that the individual species have short ranges iu time. They thus afford help to the geologist endeavouring to correlate scattered or isolated cavern * Kellogg, Univ. Cal. Publ. Zool., 21, p. 245, 1922. ' HiNTON, Ann. Mag. N.H., [9], 11, p. 162, 1923. * WiNGE, Jordfundne og nulevende Gnavere fra Lagoa Santa, p. 601. * WiNGE, Vidensk. Meddel. naturh. Foren. Kjobenhavn, 1881, 1882, p. 24 (footnote). » Schwann, P.Z.S., 1906, p. 110. 126 MICROTIN^ and fissure deposits with others to which ordinary stratigraphical methods can be applied. The succession, beginning with the oldest, would seem to be as follows :— UPPER PLIOCENE OR CROMERIAN. a. Norwich Crag, Weybourne Crag, and a stratum unknown m situ, but represented by the " clay pebbles " occurring in the Lower Freshwater Bed of the Norfolk Coast. The voles known from these deposits are all species of Mimomys, namely : '^Mimomys plioccenicus Forsyth Major. ^Mimomys reidi Hinton. \Mimomys newtoni Forsyth Major. b. Shelly Crag at East Runton. The species known from this deposit are ; — "f Mimomys plioccBfiicus Forsyth Major. j; Mimomys intermedins Newton. ^Mimotnys savini Hinton. c. Upper Freshwater Bed at West Runton. Four genera and ten species are known from this horizon : — "fMimomys intermedius Newton. ^Mimomys savini Hinton. ^Mimomys majori Hinton. ^Evotomys sp. (E. glareolus group). "fPitymys gregaloides Hinton. ^Pitymys arvaloides Hinton. ^Microtus arvalinus Hinton. "fMicrotus nivaloides Forsyth Major. '\Microtus nivalinus Hinton. ^Microtus ratticepoides Hinton. Among other characteristic mammals of the Cromerian deposits may be mentioned Macaca, MachcBrodus, Trogontherium and Hippopotamus. Pitymys gregaloides and Machcerodus have both been found in Kent's Cavern, and their presence suggests that among the many deposits of that famous cave there is a stratum of Cromerian age. PLEISTOCENE. a. High Terrace of the Thames. Remains of three species have been obtained from a small section in the High Terrace gravel at Ingress Vale, near Greenhithe, Kent. These are : — "f Mimomys cantianus Hinton. '\Evotomys sp. {E. glareolus group). Microtus or Pitymys sp. BRITISH FOSSIL SPECIES 127 Fig. 6. Fia. 64. — -Lower molars of Cromerian species of Pilymys^; all from the Upper Freshwater Bed at West Runton with the exception of fig. 12. 1. Pilymys arvaloides Hinton. Left jWj and m^ {Savin Coll., No. 656.5). 2. P. arvaloides. Left vi^ {Savin Coll., No. 8.38). 3. P. arvaloides. Left m^ {Savin Coll., No. 8.35). 4. P. arvaloides. Left m^ {Savin Coll., No. 8.37). 5. P. arvaloides. Left 7rti {Savin Coll., No. 656.6). 6. P. arvaloides. Left Wj and jWj {Savin Coll., No. 491.6). 7. P. arvaloides. Left Wj and m^ {Savin Coll., No. 8.28). 8. P. arvaloides. Right wij with vestige of cusp 2 {Savin Coll., No. 656.2). 9. P. arvaloides. Right m^ (Savin Coll., No. 8.28). 10. P. arvaloides. Left wij. 11. P. gregaloides Hinton. Left Wj, in^, type (B.M., No. 12345; Hinton 12 15. P. gregaloides. 16. P. gregaloides. 17. P. gregaloides. 18. P. gregaloides. Right wij, Kent's Cavern (B.M., No. 15084a; McEnery Coll. P. gregaloides. Coll.). 13. Pilymys sp. Left m, {Savin Coll., No. 8.30). 14. P. gregaloides. Left Wj {White Coll.). Right Ttii {Savin Coll., No. 8.24). Left w,. Left TOi {Savin Coll., No. 8.26). Left nil *f"l "*2 {Savin Coll., No. 491.5). ' For diagnoses of these species see Hinton, Ann. Mag. N.H., [9], 12, p. 541, 1923 ; a full account is given in Vol. II of this Monograph. 128 MICROTIN^ Fig. 27. Fig.28. Fig.29. Fig. 65. — Lower molars of Mkrotus from the Upper Freshwater Bed i (Cromerian). (For explanation see opposite page.) 1 For diagnoses of these species see Hinton, Ann. Mag. N.H., [9], 12, p. 541, 1923 ; a full account is given in Vol. II of this Monograph. BRITISH FOSSIL SPECIES 129 Explanation ok Fig. 65. All the specimens are from the Upper Freshwater Bed at West Runton. 1. Microtuti arvalinu/i Hinton. Right m-^, young, with ephemeral com- plications; la basal view. 2. M. arvalinus. Left wij. 3. M. arvalinus. Right m^ and m^ [Savin Coll., No. Gi4.8). 4. M. arvalinus. Right r/ij (Savin Coll., No. 644.8). 5. 31. arvalinus. Right Wj {Savin Coll., No. 654.1). 0. M. arvalinus. Left Wj. 7. M. arvalinus. Left m^ {Savin Coll., No. 643.10). 8. M. arvalinus. Left m^ {Savin Coll., No. 643.0). 9. AI. arvalinus. Right r«i {Savin Coll., No. 7.25). 10. M. arvalinus. Left ?«j {Savin Coll., No. 643.1). 11. iM. arvalinus. Left m^ {Savin Coll., No. 643.2). 12. M. arvalinus. Right wtj. 13. M. arvalinus. Left m^ {Savin Coll., No. 643.11). 14. M. arvalinu.s. Right m^. 15. M. nivaloidcs Forsyth Major. Right »»i {Savin Coll., No. 12.26). 16. 31. nivaloides. Right m^ {Savin Coll., No. 654.5); 16a part of basal view. 17. 31. nivaloides. Left m^. 18. 3J. nivaloides. Right m^. 19. 31. nivaloides. Left nii {Savin Coll., No. 7.11). 20. M. nivaloides. Left Wj {Savin Coll., No. 643.6). 21. 31. nivaloides. Left Wj. 22. M. nivalinus Hinton. Left Wj. 23. 31. nivaloides. Left m^ {Savin Coll., No. 643.4). 24. 31. nivalinus ? Left m^ and in^ {Savin Coll., No. 8.17). 25. 31. nivalinus. Right Wj {Savin Coll., No. 8.18). 26. 31. nivalinus. Right »«i {Savin Coll., No. 643.5). 27. 31. nivalinus ? Left ?«i and Wj {Savin Coll., No. 643.12). 28. 3fl. nivalinus. Left m^, type {Savin Coll., No. 8.16). 29. 31. ralticejioides Hinton. Left m^ and 7«2 {Savin Coll., No. 653.3). Trogo)itJierium aud the remaiiis of many other species indicat- ing a rich mammalian and molluscan fauna occur also at Ingress Vale. The whole assemblage closely resembles that known from Cromerian horizons and suggests that there is no great difEerence of age between the portion of the High Terrace represented at Ingress Vale and the Upper Freshwater Bed at West Runton. h. Early Middle Terrace of the Thames (typical locality, Grays Thurrock, Essex). Three species are known viz. : — "fArvicola prceceptor Hinton. \Evotomys sp. {E. glareolus group). "fMicrotus agrestoides Hinton. Macaca, Hippopolamus, Rhinoceros megarJiinus and Elephas anliqmts are characteristic survivals from the Upjser Pliocene occurring in this horizon. The species of Arvicola is not closely related to modern species of that genus ; it seems to be a direct descendant from one of the species of Mimoinys occurring in the Upper Freshwater Bed. Very similar forms are found in the latest " Cromerian " horizons as in the so-called Upper Fresh- V.L. K Fig. G6.— Cheek-teeth of recent and British fossil members of the Microtus nivalis group (Hinton, Proc. Geol. Assoc, 20, p. 39, PI. 1, IMl)- (For explanation see opposite page.) BRITISH FOSSIL SPECIES 131 Explanation of Fig. 66. I, 2, 3 and 4. M. nivalis Martins. Recent; Apennines, Italy (Dr. Forsyth Major). 5. M. agrestoides Hinton. Right nii. Pleistocene; Grays. 6-14. 31. nivalis group. Pleistocene; Middle Terrace deposits, Crajrford. 15. 31. nivalis group. Middle Terrace, Wicltham. 16-23. 31. nivalis group. Pleistocene; Clevedon Cave. 24-27. M. malei Hinton. Pleistocene; Clevedon Cave. 28. Microtus sp. Clevedon Cave. 29. 3Iicrotu8 sp. Middle Terrace, Crayford. water Bed of Bacton. These facts and those mentioned above in connection with the High Terrace appear to show that the Cromerian beds on the one hand, and the High and early Middle Terrace deposits of the Thames on the other, are in part con- temporary, and may in fact be regarded as parts of one great whole. Confirmation of this view may be found at Piltdown, where in a deposit that is evidently the stratigraphical equivalent of the High Terrace of the Thames, the remains of a fauna still more ancient than that represented at Ingress Vale have been found. Eoanthropus, Mastodon, and Stegodon are genera which might well carry us back to Norwich Crag times. These facts and deductions have a not unimportant bearing upon the question as to the date of the major glaciation of Britain ; but that is a subject which I have discussed elsewhere.^ c. Late Middle Terrace of the Thames (typical locality, Crayford and Erith). The following species are known : — "fDicrostonyx gulielmi Sanford. *Lemmus lemmus Linnseus. * Microtus nivalis group. ifMicrotus malei Hinton (and allied forms). *Microtus ratticeps Keyserling and Blasius. Between Early and Late Middle Terrace times the Microtinse, together with most of the other mammals of Britain, were com- pletely changed. Old forms, surviving in a more or less modified condition from the Pliocene period, now became extinct, and were replaced by an entirely new assemblage. Microtine remains occur ^ Hinton and Kennard, " The Relative Ages of the Stone Implements of the Lower Thames Valley," Proc. Geol. Assoc, 19, p. 76, 1905. Hinton and Kennakd, " Contributions to the Pleistocene Geology of the Thames Valley. 1. The Grays Thurrock Area." Part II, Essex Naturalist, 15, p. 56, 1907. Hinton, " Preliminary Account of the British Fossil Voles and Lem- mings," Proc. Geol. Assoc, 21 p. 489, 1910. Hinton, " Rivers and Lakes," London, Sheldon Press, 1924, pp. 49-177. Hinton, " The Pleistocene Mammalia of the British Isles and their bearing upon the date of the Glacial Period," British Association, Southampton, 1925. (In the press. Yorkshire Geological Society's Pro- "" •) 132 MICROTIN^ Yjg. fi". — Ix)wer molars of British Pleistocene species of Microtus (Later Middle Terrace forms). (For explanation see opposite page.) BRITISH FOSSIL SPECIES 133 Explanation of Fig. 67. Microtus sp. (allied tn M. malei). Figs. I, 3, 4, 5, 7, 8, 9, 12, 14, 21, 26, from the aevednn Cave. Figs. 2, 6, 10, 11, 13, 15-20, 22, 25, 27, 28 and 31, from the Later Middle Terrace deposits of Grayford. Microtia ratticeps K. and Blasius. Clevedon Cave. (Fig. 29.) M. ratticeps. Ightham Fissures. (Fig. 30.) abundantly in later Middle Terrace horizons, not only at Grayford and Eritli but in other deposits which I regard as contemporary {e.g., the Clevedon Cave and Banwell Cave). Just how many species should be recognized among the voles listed above as members of the " M. nivalis group " and " M. malei and allies" is difficult to say ; but these forms are strongly marked and especially characteristic of this horizon. I have examined many thousands of Microtine fossils from a great many British deposits, but have never found these peculiar late Middle Terrace forms in association with any species other than those mentioned in the list above. d. Ightham Fissure Stage. The species are : — '\Dicrostonyx henseli Hinton. *Letmnus lemmus Linnaeus. "fEvotomys harrisoni Hinton. ^Evotomys kennardi Hinton. \Arvicola abbotti Hinton. *Microtus ratticeps Keyserling and Blasius. "^Microtus anglicus Hinton. *Microtus arvalis Pallas. '\Microtus corneri Hinton. Microtus agrestis Linnaeus. In deposits of this age Dicrostonyx henseli, Microtus anglicus, M. arvalis, and M. corneri are especially abundant, whereas Lemmus lemmus and Microtus ratticeps are rare. Most of the cavern deposits of Britain are intermediate in age between Late Middle Terrace and the Ightham Fissure stages. In these caves one finds Lemmus lemmus and Microtus ratticeps to be more abundant than at Ightham, and Dicrostonyx henseli is frequently replaced by the older form D. gulielmi. Rarely, as at Merlin's Cave, in the Wye Valley, and at the Langwith Cave in Derbyshire, the two species of Dicrostonyx are associated. Of the two species of Evotomys found at Ightham, one {E. harrisoni) is a representative of the living E. glareolus, the other {E. kennardi) is a member of the E. nageri group and therefore probably the forerunner of the peculiar forms which now inhabit some of the smaller islands around the British coast {E. skomerensis, E. ccBsarius, E. alstoni and E. erica). Arvicola abbotti is a widely 134 MICROTIN^ Fig. 68.-Lower molars of fossil and living species of the subgenus Stenocranius. ,,|g.M^Je™.«.^ from the Ightham Fissures (Late Pleistocene). Fia 11. Uphill Cave, Weston-super-Mare it/ I.) «ia«.Lnic«. Biichner. R^-^*- (^f^. ^^^^ '^'^ ' Jf. Stenocmnitw) sp. Recent; Mongolia. (Fig. 23.) M. (S.) angustus Thomas. Recent. (i?ig. ^7.) BRITISH FOSSIL SPECIES 135 distributed and characteristic Late Pleistocene form; it is specialised for fossorial habits, and is no doubt nearly related to the living members of the A. scherman group. Of the species of Microlus, M. corneri is of special interest as being the forerunner of the peculiar M. orcadensis and its representative in the Channel Islands, M. sarnius. e. Third Terrace of the Thames and its tributaries (typical development in the Lea Valley at Ponder's End and Angel Road. Three species are at present known to occur : — "fDicrostonyx henseli Hinton. fMicrotus anglicus Hinton. *Microtus arvalis group. In deposits of this stage remains of Dicrostonyx henseli occur in great abundance, all other species being extremely rare or absent. Since the publication of Mr. S. Hazzledine Warren's accounts of the Third Terrace in the Lea Valley ^ it has been generally recognized that the terrace in question represents a cold period. The discovery of a cold fauna and flora at that horizon was predicted by me ^ ; and in my opinion the Third Terrace marks the opening of the major glaciation of this country. In surveying the Pliocene and Pleistocene succession in Britain it is not until we reach late Pleistocene times that we find any evidence at all of a " cold fauna or flora " or any real evidence of glaciation ; it is not possible to frame a glacial theory that will fit all the facts, be they physiographical, stratigraphical, or palaeontological, unless we are prepared to place the " major glaciation " at this comparatively late moment. Many deductions as to former climatic conditions have been and no doubt will continue to be made from the presence or absence of particular species or groups of species in successive geological horizons. Such deductions should, however, be made with the greatest caution. In my opinion no reliable evidence of changes of climate is afforded by the fossil mammals of countries now temperate. Very few species are unable to exist in temperate conditions ; their presence or absence depends upon access, food, shelter, and competition, rather than upon climate. Among Microtinae, Microlus nivalis, the Snow Vole, buried for ten months in a year under Alpine snows, and Dicrostonyx, the " warmth- hater " of Hcnsel, have been favourite subjects for geological comment and speculation. The presence of their remains in abundance on the plains of temperate Europe has been regarded again and again as a strong proof of former severe climatic con- ditions. But M. nivalis still lingers on in the hot lowlands of southern France, and Dicrostonyx lives in the by no means Arctic » S. Hazzledine Warebn, Q.J.G.S., 68, p. 213, 1912; and 71, p. 164, 1916. ' HiXTON, Proc. Geol. Assoc, 21, p. 503, 1910. 136 MICROTINiE climate of Unalaska. It may be worth mentioning that a specimen of D. Tiudsonius from Labrador lived in the Museum rather close to a radiator from early September to Christmas, that far from hating warmth it took the utmost care to keep warm, and that its death was due not to the subtropical atmosphere of the Upper Mammal gallery, but to a surfeit of Christmas dainties provided by mistaken kindness. N.B.- — ^In the lists of species given above, as elsewhere in this Monograph, a dagger (f) precedes the name of each extinct species ; an asterisk (*) indicates that the species is extinct in Britain though living elsewhere. II. SYSTEMATIC REVISION OF THE MICROTIN^ Group: LEMMI. Lower incisor short, its alveolus not extending backwards beyond 7n^, lingual to molars throughout. Genus : 1. DICROSTONYX Gloger. 1779. Mas Pallas, Nov. Spec. Quadr. Glires Ord., p. 77 (in part). 1795. Lenimus Link, Beytr. Naturgesch., 1, pt. 2, p. 74 (in part). 1808. Sphala.t (misprint for Spalax) Tiedemann, Zoologie, p. 476. Not of Gueldenstaedt, 1770. 1811. Myodes Pallas, Zoogr. Rosso-Asiat., 1, p. 173 (in part). 1813. Brachyurus Fischer, Zoognosia, ed. 3, 1, pp. 14, 24; 3, 1814, p. 5.5 (in part). 1817. Arvicola Cuvier, Regne Anim., 1, p. 207 (in part). Not of Lacepede, 1801. 1817. Oeorychus Cuvier, Regne Anim., 1, p. 207; Richardson, 1829. Not of Illiger, 1811. 1827. Hipudceas Lesson, Man. Mamm., p. 277 (in part). 1829. Hypydceus Fischer, Synopsis Mamm., p. 299 (misquoting Lesson). 1830. Cuniculus Wagler, Nat. Syst. Amphib., p. 21. Not of Brisson, 1762. 1841. Dicroslonyx Gloger, Hand- und Hilfsbuch Naturgesch., 1, pp. xxxi, 97. 1845. Lemims Schinz, Synopsis Mamm., 2, p. 255 (in part; misprint for Lemmus). 1845. Hypudceus Schinz, Synopsis Mamm., p. 250. Not of Illiger, 1811. 1854. Myolemmus Pomel, Catal. Method., p. 27; based upon an unidentifiable species, " Arvicola ambiguus " Pomel, occiu'ring in the Pleistocene of France. 1855. Misolhermus Hensel, Zeitschr. deutsch geol. Gesellsch., 7, p. 492; based upon Myodes torquatus Pallas. 1881. Borioikon Poljakov, Mem. Acad. Imp. Sci. St. Petersbourg, 39, Supplement, p. 34; type Myodes torquatus Pallas. 1887. Cuniculatus Nelson, Report Nat. Hist. Coll. Alaska, p. 278 (misprint for Cuniculus). 1897. Tylonyx Schulze, Mamm. Europsea in Helios, Abiiandl. u. Vortriige Gesammtgeb. Naturw. Berlin, 14, p. 83. Genotype. — An American species, probably Mus hudsonius Pallas. 138 MICROTINiE Distribution and Range in time. — Circumpolar. In the Old World it is not known to occur in Lapland, but from the eastern shore of the White Sea its range extends more or less continuously eastwards to the Behring Strait. In this region apparently it does not now come south of latitude 68° N. ; northwards it is known to reach 82° N. latitude, inhabiting Spitzbergen,^ Novaya Zemlya and the New Siberian Islands. Fossil remains of at least two species occur abundantly in the late Pleistocene deposits of Western and Central Europe, indicating that the range of the genus was formerly far more extensive in the Old World than at present, extending southwards and westwards to Ireland, southern England, southern France and the Swiss Alps. Remains of the genus are also known from the late Pleistocene of Central Asia, indicating that its range in that continent formerly extended southwards at least to the Altai Mountains. In America the genus ranges through the Labrador Peninsula, the whole of Arctic America from the western shore of Hudson's Bay to Alaska and the islands of St. Lawrence and Unalaska, the Arctic Archipelago, and northern and eastern Greenland. No fossil remains have, as yet, been found in the New World. Characters. — Medium-sized Microtinse, sharply distinguished from all other genera by their highly modified external characters, remarkable seasonal changes in the pelage, colour, and fore-claws, peculiar skulls and characteristic cheek-teeth. Size medium, hind-foot 15-20 mm., condylo-basal length of skull 29-32 mm. Fur dense, long, very soft and silky. Eyes moderate, larger than in Lemmus. Ears very small, each reduced to a low fold of naked integument encirchng the meatus and entirely concealed beneath the fur ; pre-auricular ear-tufts, formed by hairs arising from the region immediately in front of the meatus, well marked, capable of motion independently of the rest of the fur and serving completely to close the meatus when laid back by muscular contraction. ^ Tail short and cylindrical, its vertebrae about equal to the hind-foot in length, densely clothed with long adpressed bristles, forming a true terminal pencil which often somewhat exceeds the vertebral 1 No material has been seen from Spitzbergen, but Arctic Lemmings are said by Heuglin to be carried there occasionally by drift ice. Parry found the skeleton of one upon an ice-floe to the north of Spitzbergen in lat. 81° | N. (Appendix to " Narrative," 1828, p. 190.) Heuglin says : " In der Advent-Bai im Is-fjord stiess ich iibrigens an giinstigen, sommerlich gelegenen Oertlichkeiten ofter auf Lemmingbaue und unser Harpunier versicherte mich, in derselben Gegend diese Thiers ausgegraben zu haben " (Reisen n. d. Nordpolarmeer, 1870-1871, Theil 3, p. 8, 1874). ^ The character and independent motion of these ear-tufts have been described by Heuglin (Reisen n. d. Nordpolarmeer, 1870-1871, Theil 3, p. 10, 1874), who had opportunities of examining living examples of D. torquatus. Recently, thanks to the generosity of Mr. E. Else, I have been able to observe a living specimen of D. hudsonius from Labrador, and I am able to confirm HeugUn's statements. DICROSTONYX 139 portion of the tail in length. Hands and feet short and broad, each with five digits, the palms and soles densely furred. In the hand, the thumb is reduced to a small vestige which bears, however, a minute, flattened nail; on the palmar side of the thumb is a large polhcal tubercle, two or three times as large as the thumb, covered with horny integument ; the four outer digits are very short and subequal in length, digits II and V being but slightly shorter than digits III and IV. The lateral fingers (II and V) bear long, slender, laterally compressed claws which, apart from their great size and somewhat unusual vertical thickness, do not differ much from those of ordinary lemmings. The claws of the two central digits (III and IV) are extraordinary structures subject to a remarkable seasonal change ; in young specimens and in adults in full summer pelage, they do not differ greatly from those of other lemmings ; but in winter they are very large, sometimes exceeding half an inch in length, and have the appearance of being double, each of the two fingers seeming to bear two enormous claws, one on top of the other, which are separated at their tips by a more or less deep notch. When these differences in the claws first attracted attention, Pallas and other early observers, working with scanty material, were not unnaturally inclined to attribute them to sex; but Middendorff's researches in Taimvrland and in the museums of London and Munich showed that the difference was not sexual, and he was led to suspect that in some way the variation in the form and size of the claws was correlated with the seasonal pelage changes, of which he gave an excellent account.^ Coues in 1879 was the first to demonstrate exactly what takes place, and he described the process as follows ^ : — " In spring and early summer, these claws [III and IV] do not appear very different from those of Myodes {Lemmus], though averaging larger, more bulbous at base underneath, with the terminal portion slenderer, straighter, and sharper. This bulbous portion underneath grows out simultaneously with increase in length and amount of curvature of the main portion of the claw, until it equals or even exceeds the length of the latter, and is quite as stout, or even stouter, being somewhat broad and pad-like. At this period, it runs the whole length of the claw, from which it is separated by a groove along the sides, and by a notch at the end, both of varying depth. The claw then looks like nearly two claws, one underneath the other. The pad would then seem to gradually sever its connection with the main claw by progressive increase of the constriction marked by the lateral groove and terminal notch, as well as by loosening from the base, when it appears like an e.xcrescence ; it is finally lost. Thus the process . appears to be a periodical one, like the shedding of the horns of ruminants, and not continually progressive with age; and would seem to be connected witli the particularly fossorial habits of the quasi-hibernating animal that digs galleries underground in which to ■ MiDDENDORFP, Sibirischc Reise, 2, Th. 2, p. 93, 1853. • CouES, Monographs N. Amer. Rodentia, Muridae, 1879, p. 248. 140 MICROTIN^ reside during the cold season, as compared with its freer and more active mode of life in summer." Hind-feet remarkably broad, the width across the bases of the digits equal to about one-half the length, all the toes being rather short; hallux well developed, scarcely shorter than digit V ; , digits II, III and IV longer and subequal, III being, however, slightly longer than its neighbours ; all, including the hallux, bear long, slender, curved, and sharply pointed claws, which equal or surpass their digits in length. In both hands and feet the claws are covered and protected by long hair growing from the dorsum and sides of each paw and its digits; these hairs always project beyond the claw tips, the shorter summer hairs being replaced by a longer growth in the winter, which keeps pace with the increasing length of the claws. Palms and especially the soles densely clad with long crisp hair, which on each sole forms a stiff curling brush quite comparable with that developed upon the foot of a hare. When the hairy covering is clipped away the palms are seen to be quite smooth, bearing, apart from the great pollical tubercle, no trace of the pads present in normal Microtinse. Normally too, judging from a specimen from Discovery Bay specially examined by me and from two studied by TuUberg,^ no trace of the pads remains upon the sole ; but Miller ^ states that he found " several minute, faintly developed tubercles near the base of the toes." Mammae, in adult females, 2 — 2 = 8. Colour in summer brownish or greyish above, whitish or yellowish below; often with brilliant rufous suffusions upon the fore-quarters and throat. In winter pure white above and below, Dicrostonyx being the only genus among Simplicidentata that shows the phenomenon of winter whitening completely developed.^ Young like adults in summer pelage, but duller, and with a dark stripe along the spine. This stripe, in some forms, persists also in the adult summer pelage. Hair, with the exception of the special hairs (which are whitish throughout) clothing the tail and feet, everywhere with slaty bases, the colour, both in summer and in winter pelage, being produced by the terminal parts of the hairs. In the summer coat the hairs are shorter, 15-18 mm. long on the back, and their terminal portions have from one to three diversely coloured annulations, which vary both in number and hue with the species or subspecies and with their precise situation upon the body. In winter the hairs are longer, about 25 mm. on the back, with pure white tips. The change is made by two moults, one in autumn, the other in spring; it is brought about by the rapid growth of new hairs and the more gradual shedding '^ TuLLBERG, Ueber das System der Nagetiere, p. 257, 1899. " Miller, N. Amer. Fauna, No. 12, 1896, p. 37. ^ In Lemmus obensis novosibiricus winter whitening takes place also i (see p. 203). \ DICROSTONYX 141 of the old ones. In the autumn moult the whitening usually starts low down on the flanks and extends gradually upwards and forwards to the crown of the head, whereas in the spring moult the coloured coat of summer first appears upon the head and shoulders and gradually extends backwards and downwards. The process is nevertheless subject to a great deal of individual variation.^ Skull (Fig. 69) rather strongly built, moderately broad and but little depressed. Rostrum, in correlation with the rather weak upper incisors, light and slender, the anterior palatal foramina rather large, the diastema and nasals rather long. Nasals extending backwards about as far as the ascending branches of the premaxillaries, ending at a point usually a httle in front of a line connecting the anterior margins of the orbits. Zygomatic arches given off squarely from the sides of the rostrum, the greatest zygomatic breadth faUing in the anterior or maxillary parts of the arches and amounting to between 62 and 70% of the condylo-basal length. Upper border of jugal raised into a moderately high convex crest for the insertion of the temporal fascia. Temporal ridges salient throughout; in the moderately constricted interorbital region they are usually persistently separated by a longitudinal sulcus, which becomes, however, narrower and deeper with age until, in some exceptional indi- viduals, the groove may be interrupted at one point by the ridges coming into actual contact with each other. Posteriorly the temporal ridges traverse the parietals and squamosals, but their course is slightly below the level of the lateral borders of the interparietal. Anteriorly the parietals are widely and shallowly emarginated by the intertemporal portion of the coronal suture. * An interesting experiment made by Ross,* in arctic seas, nearly a century ago is worthy of notice. He kept an individual in its summer coat alive in his cabin through the winter months, and on February 1 it still retained the summer pelage. On that date Ross removed the animal from the cabin to the deck and there exposed it to the rigours of the arctic winter. Within a week, when the experiment was terminated by the death of the captive, the coat turned wholly white except in a relatively small area over the spine. By clipping og the white hair-tips Ross found that he could restore to the animal the appearance it possessed while in summer pelage, and he therefore thought that the change had taken place by a bleaching of the tips of the hairs. The true interpretation is doubtless, however, that given by MiddendorfE,t viz., that the hairs of the winter coat were all present though concealed under the tips of the summer hairs while the animal remained in the cabin, the unnatural warmth of that place retarding their growth and so delaying the shedding of the old hairs ; that the sudden exposure accelerated the growth of the new hairs which rapidly attained their full length and concealed the far shorter hairs of the old coat; and that finally sufficient time did not elapse between the exposure to cold and the death of the animal to permit of the old coat being cast, and therefore Ross was able to bring that old coat once more to the surface by cutting off the ends of the newer and longer hairs. * Ross, App. Narr. Second Voyage, 1835, Nat. Hist., p. xiii. t MiDDENDORFF, Sibirische Reise, 2, Th. 2, p. 91, 1853. 142 MICROTIN^ In the newborn animal the distance between the squamosals in front is roughly equal to twice the interorbital breadth ; in the adult the space between these bones is diminished, being no greater and often less than the interorbital breadth. In the Fig. 69. — Dicrostonyx grcenlandicus Traill. Dorsal, ventral, and lateral views (enlarged) of the skull. The small out- line is a natural-sized representation of the skull in dorsal view. adult each squamosal develops a strong, prominent, peg-like post-orbital process, which gives origin to a tendinous portion of the temporal muscle. Posterior portions of squamosals very broad, forming practically the whole of the large supratympanic fossse; supratympanic fenestra very small. Inferior portions of lambdoid crest, i.e., all below the supratympanic fenestra DICROSTONYX 143 on each side, formed by the squamosals alone, the lateral processes of the supraoccipital being short. Interparietal of normal shape, wider than long ; in some forms it articulates with the squamosals, but in others it is separated from those bones by narrow tongues of the parietals which pass back on either side to articulate with the supraoccipital. The cheek-tooth rows diverge slightly behind ; Fig. 70. — Dicrostonyx henseli Hinton. Dorsal, ventral, and lateral views (enlarged) of the type skull from the late Pleistocene fissure deposit at Ightham, near Sevenoaks, Kent (B.M., No. M.l 1,803, Geol. Dept.). The small outline is a natural-sized representation of the skull in dorsal view. the alveolar capsules, in consequence of the great height of the teeth, rise up in the floor of the orbit and in the sphenorbital fissure ; in correlation with this and with the breadth of the teeth, the presphenoid is reduced in most species to a slender bar. Palatal structure differing from that of Microtus chiefly in the extension further forwards of the mesopterj'goid fossa and in the shortness and free termination of the post-palatal median septum, which is here represented merely by a short median Fig. 71. — Dicrostonyx. Crown views of upper molars : 1. Dicrostonyx lorquatus Pallas (Miller Coll., B.M., No. 7.7.7.3621). Eight )?;•, m', and m'. 2. Same specimen. Left m' and »?i'. 3. Dicrostonyx groe^ilandicus Traill. Discovery Bay (B.M., Xo. 77.8.6.5a). Eight m', m', and ni'. 4. Dicrostonyx giilielmi Sanford. Pleistocene; Hutton Cave, Somersetshirs. Upper molars, right and left, of the typical skull (Taunton Museuiii), 5. Dicrostonyx gulielmi Sanford. Pleistocene ; Kesh Caves, Co. Sligo. Eight m} and m' (Dublin Museum, C.K. 12). 6. Dicrostcmyx gulielmi Sanford. Pleistocene; Dog Holes, Warton Crag, Lancashire. Eight ?«' and m' (B.M. ; presented by J. Wilfrid Jacksmi). 7. Dicrostonyx hudsmmis Pallas. Labrador. Lett ni", m', and m' (B.M., No. 61.2.4.1). 8. Dicrostonyx henseWRiniaa. Pleistocene ; Ightham Fissures, Kent. Left »i', m% and m'. 9. Dicrostonyx henstli Hinton. Pleistocene; Doneraile Caves. 9. Left m', m', and »«' ; a. right m' ; b. and c. right m' (Dublin Museum'). 10. Dicrostoni/x torquatus ungulatus v. Baer. Novaya Zemlya. Eight m', m*, and m' of a young individual (B.M., No. 80.3.29.2). Fig. 72. — Dkrostonyx. Crown views of lower molars : — • 1. Dicrostonyx grcenlandicus. Discovery Bay. Right tn^, rti^, and tn^ (B.M., No. 77.8.6.5a). 2. Dicrostonyx guUelmi Sanford. Pleistocene ; Langwith Cave, Derby- sliire. Left m^, wij, and m.^. 3. Dicrostonyx gulielmi Sanford. Pleistocene; Edenvale Caves, Co, Cork. Right Wj, m^, and m^ {Dublin Museum). 4. Dicrostonyx hudsonius Pallas. Labrador. Left m^, m^, and /n. (B.M., No. (51.2.4.1). 5. Dicrostonyx hudsonius Pallas. 61.2.4.2). 6. Dicrostonyx giilielmi Sanford. Labrador. Right m^ {B.M., No. Pleistocene; Kesh Caves, Co. Sligo, Ireland. Left ?nj (Dublin Museum). 7. Dicrostonyx hensrli Hinton. Pleistocene; Ightham Fissures, Kent. Left m^, m^, and m^. 8. Dicrostonyx henseli Hinton. Pleistocene; Doneraile Caves, Co. Clare, Ireland. Left tn^, m^. and m3 (Dublin Museum). 9. TJzcro.f/oH ?/.r sp. Pleistocene; England. Right m 3 showing abnormal reduction by insulation of the first inner infold. 10. Dicrostonyx torquatus ungulatus v. Baer. Novaya Zemlya. Right »«i, 7«2, and r«3 of young individual (B.M ., No. 80.3.29.2). 146 MICROTIN^ spine. Pterygoid fossae deep, their floors distinctly dorsal to the ventral surface of the basisphenoid. Auditory bullae of small or medium size ; middle ear filled with a rather dense sponge of bone ; canaliculus tympanicus completely ossified ; tegmen tympani articulating as iisual with the squamosal ; mastoid portion slightly but noticeably inflated. Basioccipital moderately broad in front. In consequence of the shortness and wholly lingual coiirse of the lower incisors, the mandible, like that of other lemmings, differs from that of the voles in the stoutness of its horizontal, and slenderness of its ascending rami. In adults the lower incisor on each side terminates near the hinder edge of m^; but in the newborn it does not pass m^. In each ramus the anterior border of the coronoid process rises above the molar level at a point opposite the middle of m^. Angular processes large and of normal form. Dentition. — Upper incisors without grooves, and with normal cutting edges. Lower incisors short, traversing the lower jaw on the lingual side of the molars to terminate opposite m.g. Cheek-teeth rootless and persistently growing, with their re- entrant folds destitute of cement. When unworn with tubercular caps; when worn displaying a normal prismatic pattern. In adult stages of wear (Figs. 71, 72) all the triangles are sub- stantiaUy closed, of somewhat peculiar, transversely elongated form, the inner and outer salient angles being subequal in size in each tooth. Enamel conspicuously differentiated into thick and thin portions, forming respectively the concave and convex sides of the salient angles as in the higher members of the genus Microtus, becoming very thin at, or lacking altogether from, the tips of the sahent angles. The patterns of the upper teeth (Fig. 71) are as follows : — m^ with an anterior loop, foUowed by six alternating triangles, of which the postero-external one is much reduced, and with four salient angles on each side ; m^ with an anterior loop and five triangles (three external, two internal), of which the postero-external one is reduced, and with four outer and three inner salient angles ; m^ with an anterior loop, four closed triangles, a posterior loop formed by two confluent and more or less reduced triangles, and with four salient angles on each side. In most of the known species a vestigial fifth inner angle is developed at the hinder end of m^ and a similar vestigial fourth inner angle at the hinder end of m^ ; but in two species, the living D. hudsonius and the extinct D. henseli, these vestiges are absent (cf. Figs. 71, 4 and 71, 7) and the posterior wall of the triangle immediately in front (fourth inner angle in m^, third inner in m^) is reduced, losing its concave form and a great deal of its thick enamel. This distinction, first pointed out by HenseP in 1858 and Forsyth Major 2 in 1872, ^ Hensel, Zeitschr. deutsch geol. Gesellsch., 8, p. 280. ^ Forsyth Major, Atti Soc. Ital. Sc. Nat., 15, p. 125. DICROSTONYX 147 has been used by G. M. Allen ^ recently to divide the genus into two subgenera, viz., Dicrostonyx containing the only hving species that lacks the vestigial angles, and Misothermus comprising all the others which retain, or, as Allen thinks, have acqiiired them. In the mandible, mj consists of a posterior loop, seven closed triangles (of which four are internal, three external), and an anterior loop compounded out of at least four more or less reduced and confluent triangles (Fig. 72, i-s, lo) ; occasionally the postero-external triangle, normally blended in the anterior loop, is shut off as an eighth closed triangle ; this tooth has never less than nine dentinal spaces, and five outer and six inner salient angles ; but occasionally one of the usually ephemeral elements of the anterior loop maintains its independence and persists, forming an additional outer or inner salient angle (Fig. 72, 2) ; >/?2 has a posterior loop, four alternating triangles, and a pair of vestigial angles in front ; of the latter the outer vestige is the more reduced, and the dentine of the vestigial pair is confluent with that of the fourth triangle ; not counting the vestiges this tooth has three well-developed salient angles on each side; m^ is like Wj, but a little more reduced ; its third or antero-internal triangle is sometimes partially confluent with the fourth ; of the vestigial angles in front, the outer one is lost in many species, while in D. hudsonius both the inner and the outer vestiges are lacking. Owing to the want of material from the Old World it is impossible to say how many species of Dicrostonyx exist and to determine what status should be accorded to several of the forms currently recognized. We are indebted to G. M. Allen for a revision of the American members of the genus, and free use has been made of his work in preparing the accounts of the described forms given below. Allen arranges the species, as indicated above, in two sub- genera, Dicrostonyx and Misothermus, distinguished by the presence or absence of the postero-internal vestigial angles in m} and nfi. But this difference, although useful for discriminating between species and very constant in American and in European fossil species, is to my mind altogether too slight a foundation for subgenera, particxilarly in view of the fact that in the Old World D. torquatus the characters of m^ and trfi appear to be rather inconstant (Figs. 71, 1 and 71, 2). Whether we regard the minute postero-internal angle so frequently present in these two teeth as an ancient vestigial structure (my view) or as a rudiment of a new complication (Allen's view), it seems to be just such a character as may from time to time vanish or appear quite independently in any member of the genus; no con- clusion as to the special interrelationships of the known forms, fossil or recent, can be based upon the mere presence or absence of such a structure. 1 G. M. Allen, Bull. Mus. Comp. Zool. Harvard Coll., Cambridge, Mass., 62, p. 513. 148 MICROTIN^ 1. Dicrostonyx torquatus Pallas. (Synonymy under the subspecies.) Range. — Arctic regions of the Old World from the eastern shore of the White Sea through Russia and boreal Asia ; eastward limits of range unknown. Inhabiting Spitzbergen, Novaya Zemlya, and the New Siberian Islands. Not ranging south of lat. 68° N. Characters. — Size medium ; head and body about 130 mm. ; hind-foot 20; condylo-basal length of skull 30. Essential external characters as described under the genus. Colour, in summer, brown or grey above according to the subspecies ; the brown or typical form without a spinal stripe ; the grey form {D. t. ungulatus) with a definite dorsal stripe. Skull normal, but available material not sufficient to enable us to appreciate specific peculiarities. Upper incisors moderately heavy and much more strongly curved than in D. groenlandicus (" orthodont " instead of " proodont ") ; those of D. ruhricatus being apparently intermediate. Cheek-teeth of the more complex type; m} and m"^ usually with postero-internal vestigial angles, but these are sometimes absent (Fig. 71, i); hinder wall of the postero-internal triangle in each of the two teeth named retaining its thick enamel and usually its concave shape. Remarks. — The material before me is totally inadequate for the purpose of determining the characters of this species and its relationship to the American D. ruhricatus. la. Dicrostonyx torquatus torquatus Pallas. 1779. Mus torquatus Pallas, Nov. Spec. Quadr. Glirium Ord., pp. 77, 205, taf. xi, B. 1779. Mus lenensis Pallas, Nov. Spec. Quadr. Gilr. Ord., p. 199. 1808. Spalax torquatus Tiedemann, Zoologie, 1, p. 47. 1811. My odes torquatus Pallas, Zoogr. Rosso-Asiat., 1, p. 173 ; Midden- dorfif, Sibir. Reise, 2, Th. 2, p. 87, 1853. 1817. Arvicola (Georychvs) torquatus Cuvier, Regne Anim., 1, p. 207. 1820. Lemmus torquatus Desmarest, Mammalogie, p. 289. 1827. Hipudceiis torquatus Lesson, Man. Mamm., p. 277. 1829. HypudcBus torquatus Fischer, Synopsis Mamm., p. 298 (mis- quoting Lesson). 1844. Myodes hudsonius Middendorff, Bull. CI. physmath. Acad. Imp. Sci. St. Petersburg, 3, p. 291. Not of Pallas, 1779. 1874. Cuniculus hudsonius Coues, Proc. Acad. Nat. Sci. Philadelphia, p. 196 (part). 1877. Cuniculus torquatus Coues, Monogr. N. Amer. Rodentia, Muridse, p. 246 (part). 1896. Dicrostonyx torquatus Miller, N. Amer. Fauna, No. 12, pp. 38-40 (part). 1922.t Dicrostonyx torquatus altaicus Vinogradov, Ann. Mus. Zool. Acad. Sc. Russ., 23, p. 372; Co-types :— High Mining School Collection, No. 248/11 A. Leningrad (Gtorny Institute); seven left and five right mandibular rami; described from the Caverns of Tsharysch and Khankara Rivers, N. W. Altai, 180 kilometres from Bijsk. DICROSTONYX 149 Type. — Unknown. Type locality. — Region around the mouth of the River Obi, N.W. Siberia. Range. — That of the species, excepting the island of Novaya Zemlya. Characters. — Essential characters those of the species. General colour of adults in summer pelage. — Brownish black above, more or less brightened by the rufous and yellowish sub- terminal bands of the hairs, darkest along the spine, paling to ashy grey on the sides of the head, towards the rump and lower flanks, and upon the outer and lower parts of the limbs. No distinct dorsal stripe ; but a more or less distinct dusky streak upon the top of the head from the tip of the nose to the crown. Ear-tufts, shoulders and humeral regions, together with the foremost parts of the flanks, bright brownish-red, this colour being continued ventrally to form a more or less complete and extensive collar ; the red tint more or less distinctly interrupted by a pallid streak passing upwards from the throat on each side to the region behind the ears. Under parts, apart from the pectoral collar but including the inner surfaces of the limbs, yellowish or dirty white. Feet and tail hairs white. For external and cranial measurements, see tables at end of volume. D. t. altaicus Vinogradov, described from the Pleistocene cavern deposits of N.W. Altai, is said to be distinguished from the typical subspecies by the structure of m^, which possesses eight instead of seven closed triangles ; but this character is wholly unreliable. Vinogradov himself finds it " not quite constant " in either the fossil or the recent material before him, and if he had been able to examine the enormous amount of fossil material before me he would never have attributed systematic importance to such a detail. 1^. Dicrostonyx torquatus ungulatus v. Baer. 1841. Lemmus ungulatus von Baer, von Baer and Helmersen, Beitriige, 4, p. 283. 1853. Myodes lorqualus var. pallida Middendorff, Sibir. Reise, 2, Th. 2, p. 93 ; Heuglin, Reisen nach dem Nordpolarmeer, Tlieil, 2, frontispiece, 1873, Theil, 3, p. 6, 1874. Type.— Unknown. Type locality. — Novaya Zemlya. Range. — Known only from the type locality. Characters. — Essential characters as in the species. General colour of adidts in summer pelage. — Upper parts ashy- grey in general colour; with a distinct blackish spinal stripe, running from the root of the tail to the occiput, and continued forwards to the nose by a less definite blackish streak. Ear- tufts, shoulders, foremost parts of flanks, and breast heavily 150 MICROTINJE washed with a bright reddish-chestnut brown, much as in D. t. tor- quatus. Under parts, feet and tail whitish. For external and cranial measurements see tables at end of volume. Remarks. — The descriptions of this animal, given by von Baer, Middendorff and Heuglin, and the figure published by the latter, seem sufficient to prove its distinctness from D. t. torquatus. Pending the acquisition of modern material it may be regarded as a subspecies of D. torquatus, but its precise status is doubtful. 2. Dicrostonyx chionopses Allen. 1914. Dicrostonyx chionopcBS G. M. Allen, Proc. New England Zool. CI., 5, p. 62. Type. — Museum of Comparative Zoology, Harvard Coll. Type locality. — Nijni [Nischne] Kolymsk, near the mouth of the Kolyma River, N.E. Siberia. Range. — Known only from the type locality. Characters. — Smaller than D. torquatus (total length only 116 mm.), but with a relatively large skull (condylo-basal length 27-5 mm.). Head and body length unusually small, scarcely exceeding 100 mm. Hind-foot relatively very large, 20 mm. Summer pelage imperfectly known ; winter pelage pure white as usual. The type and only known specimen, is an adult male collected October 15, 1911, which has nearly completed its change into the winter dress. The only remnants of the summer coat remaining are a small patch of hazel on the top of the head and nape, another on the centre of the chest, and a concealed substratum of russet-tipped hairs with a few black hairs intermixed in the mid-dorsal region. Middle fore-claws already much enlarged. Skull. — The skull is described as " quite adult with the basioccipital suture solidly fused and ridged." In relation to the size of the animal it is very large; only slightly smaller in all respects than the skull of D. torquatus from the Taimyr Peninsula figured by Middendorff. Interparietal nearly rectan- gular, without an anterior median projecting point. Cheek-teeth not described, but presumably as in D. torq^iatus. For external and cranial measurements, see tables at end of volume. Remarks. — It is to be hoped that further material will be obtained. The hind-foot measurement is so large in proportion to the total length that one is almost tempted to believe that an adult skuU has by some error been allocated to an immature skin. Assuming that no such mistake can have been made, this is a form of extraordinary interest. Baird (Mamm. N. America, p. 559, 1857) says : " The N. P. Exploring Expedition, under Captain Rogers, collected specimens DICROSTONYX 151 [of ' Myodes torquatus '] on the island of Arikamtcluclii, in Behring's Straits, near the Asiatic shore." 3. Dicrostonyx rubricatus Richardson. (Synonymy under subspecies.) Range. — Arctic America from the Alaskan Peninsula eastwards along the Arctic coast to the western shores of Hudson's Bay. Represented by a subspecies upon the island of Unalaska ; and probably inhabiting the more westerly islands of the Arctic Archipelago. Characters. — Essential characters closely resembling those of D. torquatus, of which D. rubricatus is clearly the New World representative. Characters, if any, differentiating the skull from that of D. torquatus not at present satisfactorily known. Cheek- teeth of the more complex type ; m^ and m^ usually with well- developed postero-internal vestigial angles, the hind walls of their postero-internal triangles always retaining their thick enamel and concave form ; 711^ usually with a pair of anterior vestigial angles. Outward appearance, in summer, variable with the subspecies ; the typical form from Alaska making rather a near approach in coloration to D. torquatus, while the western subspecies has a peculiarly uniform reddish-grey pelage with a sharply defined black spinal stripe extending from the head to the root of the tail. Remarks. — D. rubricatus is undoubtedly very closely related to the Old World D. torquatus, and until more material representing the latter is available, the status of D. rubricatus wiU remain doubt- ful. Three well-marked subspecies are currently recognized, two inhabiting the mainland of N. America and a third confined to the island of Unalaska. 3a. Dicrostonyx rubricatus rubricatus Richardson. 1839. Arvicola rubricatus Richardson, Zool. Capt. Beechey's Voyage, 1839, p. 7. 1845. HypudiBus rubricatus Schinz, Synopsis Mamm., 2, p. 250. 1877. Cuniculus torquatus Coues, Monogr. N. Amer. Rodentia, Muridae, p. 246 (part). Not of Pallas. 1896. IHcrostony.v torquatus Miller, N. Amer. Fauna, No. 12, p. 38 (part). 1900. Dicrostonyx nelsoni Merriam, Proc. Washington Acad. Sol., 2, p. 25; MUlcr, " List," 1912, p. 207.1 1900. Dicrostonyx hmlsonius alascensis Stone, Proc. Acad. Sci. Phila- delphia, 1900, p. 37 ; Miller, " List," 1912, p. 207. i- —^ ■ • ' TlirouL;h((ut this work these abbreviations are used for : — 1. AhLi.ER, List of North American Land Mammals in the United ■States National Museum, 1911. ISmithsonian Inst. U.S. Nat. Mus. Bull., 79, 1912, pp. 1-455. 2. Mn^LEE, List of North American Recent Mammals. Smithsonian Inst. U.S. Nat. Mus. Bull., 128, 1924, pp. 1-673. 152 MICROTIN^ 1901. Dicrostonyx hudsonius nelsoni EUiot, Field Columbian Mus. Publ., Zool. Ser. 2, p. 210, fig. 48. 1905. Dicrostonyx richardsoni Macfarlane, Proc. U.S. Nat. Mus., 28, p. 736. Not of Merriam. 1919. Dicrostonyx ruhricatus G. M. Allen, Bull. Mus. Comp. Zool., Harvard Coll., Cambridge, 62, p. 518. 1914; Dicrostonyx ruhricatus rubricatus Miller, " List," 1924, p. 397.^ U Type. — Unknown ; none specified, the description based upon Collie's notes. Type locality. — Shore of Behring Strait, Alaska. Range. — Alaskan peninsula and coastal islands, eastwards in the neighbourhood of the Arctic coast of Mackenzie to Corona- tion Gulf. Characters. — Closely resembling the Old World D. torquatus torquatus in all essential respects; but adults in summer pelage rather more brilliantly coloured and with a somewhat indistinct black spinal stripe extending from the nose to the tail. Colour of adults in summer. — " Sides of the muzzle and an area about the eyes gray, due to a mixture of short hairs, some whitish, others black-tipped. Forehead from nose to the nape, black, sometimes grizzled with a few gray hairs. This mark is j continued as a narrow black median stripe to the root of the tail. Ears marked by a tuft of rusty hairs. Shoulders nearly clear i chestnut, about morocco-red of Ridgway (1912) mixed with i whitish, this colour extending back along the sides of the thorax, and blending dorsally with the grizzled whitish and blackish of , the back; hips grayish. Lower surfaces usually washed with j orange buff, but in some specimens whitish. Tail and feet ' whitish " (G. M. AUen, 1919, p. 519). Young in first pelage resemble advdts in summer coat, but lack the brilliant colours of the adults ; general colour of upper parts uniform cinnamon buff, with black dorsal stripe from the forehead to the tail. Ear-tufts black. A clear tawny patch at the shoulder. Sides and under surface washed with ochraceous buff. Skull. — Closely resembling that of D. torquatus; differing ; from that of D. hudsonius in the relatively shorter nasals and ' slightly more squarely spreading zygomatic arches. Interparietal separated laterally from the squamosal on each side by a narrow tongue sent backwards by the parietal to articulate with the supra-occipital. Cheek-teeth. — Essentially as in D. torquatus ; a postero-internal vestigial angle constantly present in tn^ and m^. For external and cranial measurements, see tables at end of volume. Remarks. — Mr. G. M. Allen deserves great credit for identifying the Alaskan Collared Lemming with Richardson's " Arvicola ruhricatus." Richardson's brief diagnosis, based upon Surgeon Collie's manuscript notes of a mouse that " burrows in the turfy * See footnote on p. 161. DICROSTONYX 153 soil ou the shores of Behriug's Straits," — once the connection is pointed out — i.s, as Mr. Allen says, " unmistakable and describes the highly coloured adult Alaskan Lemming sufficiently well." This animal is undoubtedly very closely related to D. torquatus and it is not improbable that when satisfactory material repre- senting the Old World form is acquired, it will be necessary to treat D. rubricalus as a subspecies of D. torquatus. 3i. Dicrostonyx rubricatus richardsoni Merriam. 1825. Arvicola grcenlandica Richardson, Parry's Second Voyage, App., 1825, p. 304 (part). Not of Traill. 1835. Arvicola hudsonia J. C. Ross, J. Ross's Narrative Second Voyage. App., 1835, p. xii. 1848. 'Myodes hudsonic.us Gray, P.Z.S., 1848, p. 43. 1850. Myodes groenlandicus, Gray, App. to Rae's Narrative, Exped. Arctic Sea, 1846-1847, 1850, p. 200. 1854. Georychus hudsonius, Audubon and Bachman, Quad. N. Amer., 1854, 3, p. 81 (in part), pi. 119. Not of Pallas. 1874. Cunicidus hudsonius Coues, Proc. Acad. Nat. Sci. Philadelphia, 1874, p. 196 (part). 1877. Cuiiiculus torquatus Coues, Monogr. N. Amer. Rodentia, Muridre, p. 246 (part). 1896. Dicrostonyx torquatus Miller, N. Amer. Fauna, No. 12, pp. 38-40 (part). 1900. Dicrostonyx richardsoni, Merriam, Proc. Washington Acad Sci., 2, p. 26; Miller, " List," 1912, p. 207. 1901. Dicrostonyx hudsonius richardsoni Elliot, Field Columbian Mus. Publ., Zool. Ser. 2. p. 211. 1919. Dicrostonyx rubricalus richardsoni G. M. AUen, Bull. Mus. Comp. Zool., Cambridge, 62, p. 525; Miller, " List," 1924, p. 398. Type.— U.S. Nat. Mus., No. ^i-||., Merriam Coll. ; adult male, skin and skull, collected July 1859, by W. MacTavish. Type locality. — Fort Churchill, Keewatin, west coast of Hudson's Bay. Range. — From the western shore of Hudson's Bay, westwards through Arctic America, meeting and intergrading with D. r. rubricalus in the neighbourhood of Coronation Gulf. Characters. — Adults, in summer, of a nearly uniform ruddy grey above, with a black dorsal stripe from nose to tail. Skull with rostrum relatively longer and nasals less tapering than in D. r. rubricalus. Bulh« rather larger. For external and cranial measurements, see tables at end of volume. Remarks. — As Mr. G. M. Allen remarks this '" is a very strongly marked subspecies and though closely related to the Alaskan rvbricatus, with which it intergrades to the north-westward, it parallels hudsoiius of Labrador in its dull uniform coloration, though it is not so gray as the latter." Specimens from near the type locality are widely different in external appearance from true D. rubricalus and from D. torquatus as may be gathered V.L. M 154 MICROTIN^ from the fact that Middendoi-flf, thoroughly acquainted with the Old World species, thought Audubon's very fair figure of a specimen of D. richardsoni in summer pelage to be " completely erroneous " and " not at all to be recognized." 3c. Dicrostonyx rubricatus unalascensis Merriam. 1853. Myodes torquatus Middendorff, Sibir. Reise, 2, Th. 2, p. 92. Not of Pallas, 1779. 1900. Dicrostonyx unalascensis Merriam, Proc. Washington Acad. Sci., 2, p. 25; Miller, "List," 1912, p. 208. 1901. Dicrostonyx hudsonius unalascensis Elliot, Field Columbian Mils. Publ., Zool. Ser. 2, p. 210. 1919. Dicrostonyx rubricatus unalascensis G. M. Allen, Bull. Mus. Comp. Zool., Cambridge, 62, p. 530; Miller, " List," 1924, p. 398. Type.— U.S. Nat. Mus., No. 99,622 (Biol. Surv. Coll.) ; skull lacking occiput, found in an owl pellet, July 8, 1899, by C. H. Merriam. Type locality. — Island of Unalaska, Alaska. Range. — Known only from the type locality. Characters. — " Closely related to rubricatus, from which it differs in its relatively longer and more slender rostrum, its weaker, less broadly rounded zygomata, and slightly more pro- truding incisors. External characters unknown " (G. M. Allen, 1919, p. 531). For cranial measurements, see table at end of volume. Remarks. — This lemming was first discovered by Wosnessensky about 1850. He obtained a specimen which was examined by Middendorff, who says of it : — " Es istein halbwuchsiges Weibchen in der Tracht der Jungen von Taimyr, von denen es ununter- scheidbar ist. Am llten Mai gefangen, zeigt es an der zweiten Zehe der Vorderpfote (von innen gerechuet) die hufartige Ent- wickelung des Nagels." Unalaska, situated as it is in the Pacific off the extremity of the Alaskan Peninsula in lat. 54° N., has a mild climate ; and the presence upon it of a member of a genus otherwise restricted, so far as its hving representatives are concerned, to the desolate Arctic tundra is a fact of very great interest and one which should inspire caution in the geologist drawing inferences as to past climatic conditions from the occur- rence of remains of Dicrostonyx in the Pleistocene deposits of Western Europe. For many years Wosnessensky's discovery, recorded by Middendorff and by Hensel, was doubted (Nehring, Ueber Tundren und Steppen, 1890, p. 23) ; but it has now been fully confirmed by the U.S. Biological Survey. Curiously, although twenty-nine skulls, collected from owls' pellets, have now been obtained from the island, the survey has not succeeded in procuring a specimen in the flesh; and our knowledge of the outward appearance of the animal is limited to the statement of Middendorff cited above. DICROSTONYX 155 4. Dicrostonyx exsul G. M. Allen. 1887. Ciuikulalns torqualus Nelson, Rept. Nat. Hist. Collect. Alaska, 1887, p. 278 (part). 1919. Dicrostonyx exsul G. M. Allen, Bull. Mus. Comp. Zool., Cam- bridge, 62, p. 532; Miller, " List," 1924, p. 398. Type— Mus. Comp. Zool., Harvard College, No. 11,885; adult male, skin and skull, collected June 24, 1913, by J. Dixon. Type locality. — St. Lawrence Island, Behring Sea. Range. — Only known from the type locality. Characters. — " Similar to rubricatus, but coloring much less intense, more grayish throughout. Skull with more abruptly narrowed nasals, their proximal ends bevelled sharply to a median point instead of tapering gradually; interparietal more nearly square in general outline " (G. M. Allen, 1919, p. 532). Colour of adults in summer pelage. — Pinkish grey above, blacker on the rump, with an indistinct blackish dorsal line. Each ear-patch a mixture of ochraceous buff and tawny. Throat heavily washed with tawny. For external and cranial measurements, see tables at end of volume. 5. Dicrostonyx groenlandicus Traill. 1823. Mus gra'tilandicus Traill, Scoresby's Journ. Voyage Northern Whale-fishery, p. 416. 1824. Lemrnus hudsonius Sabine, Parry's Voyage, Suppl. to App., p. clxxxviii. Not of Pallas. 1843. Mijodes groenlandicus Wagner, Schreber's Saugethiere, Suppl., 3, p. 606. 1845. Lemnus [sic] groenlandicus Schinz, Synopsis Mamm., 2, p. 256. 1850. Ml/odes hudsonius Gray, App. to Rae's Narr. Exped. Arctic Sea, 1846-1847, p. 200. 1857. Hypudceus gramlandicus Reinhardt, Rink's Gronl. geogr. stat. beskriv., p. 8. 1868. Myodes lorquatus var. granlandicus Bro\vn, Proc. Zool. Soc, 1868, p. 3.50. 1877. Myodes torqualus Feilden, Zoologist [3], 1, p. 320; Winge, Medd. cm Gronland, 21, 1902, p. 382. Not of Pallas, 1778. 1886. Cuniculus torqualus Greely, Tliree Years of Arctic Service, 2, p. 363 (part). 1896. Dicrostonyx torqualus Miller, N. Amer. Fauna, No. 12, pp. 38-40 (part). 1901. Dicroslnni/.r hudsonius Elliot, Field Columbian Mus. Publ., Zool. Ser. 2, p. 209. Not of Pallas, 1778. ISUl. Dicrostonyx hudsonius grmnlandicus .Jacobi, Abh. u. Ber. k. zool. antiirop-ethnogr. Mus. Dresden, 12, No. 4. p. 8; Miller " List," 1912, p. 207. 1919. Dicrostonyx groenlandicus G. M. Allen, Bull. Mus. Comp. Zool., Cambridge, 62, p. 533 ; Miller, " List," 1924, p. 398. Type. — Unknown. Type locality. — Jameson's Land, East Greenland, in lat. 7F N. 156 MICROTIN^ Range. — This species inhabits the maritime districts of northern and eastern Greenland, ranging from about Cape Dalton on the east coast in lat. 69° N., northwards to the limit of land (Mary- Murray and Lockwood Islands in lat. 83° N.), and thence west- wards along the northern coast to the Kane Basin and Robeson Channel, and southwards to the margin of the Humboldt Glacier. It also inhabits the countries westwards of Robeson Channel, ranging from about 83° N. lat. southwards through Grant Land, Grinnell Land, Ellesmere Land and Baffin Land, reaching its southern limits at Cape Mercy in the Cumberland Peninsula and on the northern shore of Hudson's Straits. Its western limits of range are imperfectly known, but it probably extends to the eastern shore of the Gulf of Boothia in Cockburn Land and the Melville Peninsula ; for some of the specimens obtained by Dr. Rae, in 1847, during his expedition to the region north of Repulse Bay appear to belong to this species and not to D. ruhricatus. Characters. — A rather small species (hind-foot to 16-5 mm. ; condylo-basal length of skull to 29-6 mm.). Skull and teeth essentially as in D. torquatus ; but rostrum slender and upper incisors weak and somewhat straightened. Adults in summer with grey upper parts resembling D. hudsonius in appearance; spinal stripe represented only by a dusky streak on top of the head. Essential characters as in other members of the genus, but hind-foot unusually short, varying between 13 and 16-5 mm. in six adults from Discovery Bay. Colour of adults in summer pelage. — Upper parts with a fine grizzle of grey, black and ochraceous the general effect approaching the '"mouse-gray" of Ridgway; more or less brightened by ochraceous at the ears, on the nape, shoulders, rump, and on the flanks. The prevailing grey colour is produced by hairs which have their middle thirds white, very narrow subterminal ochraceous bands, and dusky tips. The colour is locally bright- ened by broadening of the subterminal ochraceous bands and at the ear-tufts by the entire elimination of dusky tips. Spinal stripe almost obsolete ; usually represented only by a dark streak between the nose and the withers, but in one specimen (i from Discovery Bay traceable as a narrow ill-defined streak from the rostrum almost to the rump. Under parts washed with ochraceous-orange. Tail and feet white or cream-coloured. Skull rather small and lightly built, with the rostrum slender. Parietals articulating with the supraoccipital. Auditory bullse rather large and anteriorly inflated. Dentition. — Upper incisors weak, noticeably narrower than in D. ruhricatus, less strongly curved and somewhat protruding (" proodont "). Cheek-teeth essentially as in D. torquatus, but rather light ; postero-internal vestigial angles in m^ and nfi present, but usually weakly developed ; ni^ lacking an antero- external vestigial angle. DICROSTONYX 157 For external and cranial measurements, see tables at end of volume. Remarks. — This is a well-marked form, sharply differentiated from D. rubricatus and D. torquatus by its small size, grey colour and to some extent by its cranial and dental peculiarities. It is possible, as Mr. G. M. Allen has .suggested, that it intergrades with D. ruhricattis richardsoni somewhere to the north of Hudson's Bay, and some of Dr. Rae's specimens, probably those obtained in the neighbourhood of the Gulf of Boothia, seem to lend some sUght support to this suggestion ; but pending definite proof of such intergradation, it is better to accord D. grcenlandicus full specific rank. Outwardly the Greenland Lemming rather closely resembles D. hudsonius, which, however, is distinguished by its larger size, and more importantly by its peculiar cheek-teeth. 6. Dicrostonyx hudsonius Pallas. 1779. Mils hudsonius Pallas, Nov. Spec. Quad. Glires Ord., p. 208, taf. 26, fig. A, B, C. 1808. S/jalax hudsonius Tiedemann, Zoologie, 1, p. 478. 1817. Arvicola (Georyrhus) hudsonius Cuvier, Regne Anim., 1, p. 207. 1820. Lemmus hudsonius Desmarest, Mammalogie, p. 289. 1827. Hipudwus hudsonius Lesson, Man. Mamm., p. 277. 1829. Hypydceus hudsonius Fischer, Synopsis Mamm., p. 299. 1843. Myodes hudsonius Wagner, Schreber's Saugethiere, Suppl., 3, p. 604. 1845. Lemnus hudsonius Schinz, Synopsis Mamm., 2, p. 255. 1854. Georychus hudsonius Audubon and Bachman, Quad. N. Amer., 3, p. 81 (part). 1872. Myodes torquatus var. hudsonius Forsyth Major, Atti Soc. Ital. sci. nat. Milano, 15, p. 122. 1874. Cuniculus hudsonius Coues, Proc. Acad. Nat. Sci. Philadelphia, p. 196 (part). 1877. Cuniculus torquatus Coues, Monogr. N. Amer. Rodentia, Muridae, p. 246 (part). 1896. Dicrostonyx torquatus Miller, N. Amer. Fauna, No. 12, pp. 38-40 (part). 1897. Dicrostonyx hiidsonins Bangs, Proc. Biol. Soc. Washington, 11, p. 237; Miller, "List," 1912, p. 207; G. M. Allen, Bull. Mus. Comp. Zool., Cambridge, 62, 1919, p. 514; Miller, "List," 1924, p. 398. Type. — Unknown. Type locality. — Labrador; probably the east coast. Range. — This species inhabits the barren-ground area of the Labrador Peninsula, northwards from the Straits of Belle Isle on the south-east and from about the Great Whale River (in lat. 55° N.) on the west coast. It is also found on some of the small islands along the eastern side of Hudson's Bay. According to G. M. Allen it is completely isolated from D. rubricatus, the species inhabiting the western shore of Hudson's Bay, by the wooded region around James Bay. 158 MICROTINiE Characters.- — Size rather large ; hind-foot to 24 mm. ; condylo- basal length of skull to 32 mm. Variable cheek-teeth {m}, nfi, and m^ somewhat simplified ; m^ and m^ each lacking a postero- internal vestigial angle and with the posterior wall of the hinder inner triangle more or less reduced ; m^ without anterior vestigial angles. Upper parts of adults in summer pelage nearly uniform bulfy grey ; with a median blackish line from the nape to the tail, not sharply defined ; ear-tufts, sides of body and throat buff ; belly grey. Colour of adults m summer. — General tone of upper parts, including the head, varying between buify grey and clear grey, produced by the mixture of three kinds of hairs, some grey-tipped, others with dusky tips and yellowish subterminal bands, and still others longer and entirely black, the precise hue depending largely upon the development of the yellowish subterminal ring in the second sort of hairs. A black spinal line extends from the nape to the tail, but is not sharply defined. Ear on each side surrounded by an indistinct greyish patch ; while each ear-tuft is composed of tawny hairs. Throat collar represented by a wash of tawny connecting the axilla? and sometimes continued back- wards on the chest in the middle line. A paler tinge of tawny along the flanks and about the anus. A buffy spot on each side of the nose and around the eye. Tail like the back, or with a tuft of long grey hairs. The young in summer are like the adults but darker; the spinal stripe extends to the point of the nose and is more sharply defined ; the throat collar is very indistinct. Skull. — With dorsal contour rather more convex and less flattened from before backwards than in other recent species; nasals relatively long; interparietal frequently articulating laterally with the squamosals ; presphenoid rather broad. Cheek-teeth. — m^ and m'^ with the hinder wall of the postero- internal triangle in each tooth reduced, straight or convex instead of being concavely curved, its thick enamel more or less atrophied ; no postero-internal vestigial angles, m^ without any trace of the anterior vestigial angles present in all other known species (Fig. 72, 4). For external and cranial measurements, see tables at end of volume. FOSSIL SPECIES. Fossil remains of Dicrostonyx occur abundantly in the late Pleistocene deposits of Europe, and have been found in deposits of a similar age in northern and central Asia ; but hitherto none have been detected in North America. Pomel was the first to pay attention to these remains in Europe, and his excellent description ^ of some lower jaws, from the Breche de Coudes, in the Auvergne, upon which he based his » Pomel, Catalogue Methodique, 1853, p. 27. DICROSTONYX 159 Arvicola (Myolemmus) ambiguus, leaves no room for doubting that Myolemmus is a synonym of Dicroslonyx. Unfortunately the lower jaw, in this genus, yields little that is diagnostic of species, and Pomel's description contains nothing that will enable us to determine which of the fossil species should bear the name ambiguus. The description of the lower jaws is followed by an account of a skull which Pomel referred also to A. {M.) ambiguus ; but from the characters mentioned it is evident that this skull belonged to some other genus — probably Lemmus. Pomel's species '" amhiguus " therefore cannot now be identified and the name must be dropped. In 1855 Hensel ^ described a fragmentary skull from a Pleis- tocene deposit at Quedlinburg, Saxony. Possessing at that time no recent material and knowing the skull of the recent animal only from the poor figures given by Middendorff, Hensel recog- nized the close affinity of the fossil species with the living D. tor- qiialus. He also appreciated the great differences which exist between the skull and teeth of the Arctic Lemming and those of the Norwegian Lenmiing and its allies, and he separated the former as a distinct genus Misothermus from the latter, which he left in the genus Myodes (= Lemmus). A little later Hensel procured a skull of D. torquatus from the Taimyrland, and found that the two anterior upper molars {m^ and /u') had each a minute postero-internal angle which was not present in the teeth of the Quedlinburg fossil. ^ But reasoning from his experience of the variability of the posterior ends of the upper molars in voles generally, Hensel justly remarked that this minute dif?erence must be held to be a mere individual peculiarity until such time as one could examine a sufficiently large number of dentitions to prove the contrary. In 1870 Sanford^ referred a skull from the Hutton Cave, in Somersetshire, to Lemmus torquatus ; but he based a new species of Arvicola (^4. gulielmi) upon the lower jaws of the same form from the same cave. The name gulielmi is thus available and tenable for one of the British fossil species of Dicrostonyx. In 1872 Forsyth Major * found that the m^ and m- in a skull of D. hudsonius from Labrador agreed with those of the Quedlinburg skull in lacking the postero-internal vestigial angle found in each of those teeth in D. torquatus and (according to Sanford's figure) in the skull from Somersetshire too. Forsyth Major was thus led to suggest that two .species of Dicrostonyx, one agreeing in dentition with D. hudsonius, the other with D. torquatus, might have inhabited Europe during the Pleistocene period. Between 1872 and 1910 a large number of occurrences of fossil remains of Dicrostonyx in the European Pleistocene were recorded, but no ^ Hensel, Zeitschr. deutsch. geol. Ges., 7, p. 492. - Ibid., 8, 1856, p. 279. ' San-ford, Q.J.G.S., 26, 1870, p. 125, pi. viii, figs. 2 and 4. * Forsyth Major, Atti Soc. Ital. Sci. Nat. Milano, 15, p. 125. 160 MICROTIN^ one seems to have paid any attention to Forsyth Major's sugges- tion. In the latter year the present writer, ^ with the aid of excellent material, established the fact that two species have left their remains in European Pleistocene deposits and showed that neither could be identified with any of the living forms now recognized. 7. t Dicrostonyx gulielmi Sanford. 1870. Arvicola gulielmi Sanford, Q.J.G.S., 26, p. 125, pi. viii, fig. 2. 1870. Lemmns torquatus, var. Sanford, loc. cif., pi. viii, fig. 4. Not of Pallas, 1779. 1872. Myodes torquatus Forsyth Major, Atti Soc. Ital. Sci. Nat. Milan, 15, p. 124. 1880. ''.Myodes torquatus major Woldrich, Sitzungsb. d. k. Alcad. d. Wien, math. nat. CI. 82, Abth. 2, p. 25 ; based on fossil remains from a fissure deiiosit at Zuzlawitz, near Winterberg, Bohemia. 1901. Dicrostonyx torquatus Bate, Geol. Mag., [4], 8, p. 105. 1910. Dicrostonyx gulielmi Hinton, Ann. Mag. Nat. Hist., [8], 6, p. 38. Co-types. — Taunton Museum. Type locality and Horizon. — Hutton Cave, Somersetshire. Late Pleistocene. Nomenclature. — Sanford's " Arvicola gulielmi" based upon five lower jaws of Dicrostonyx from the late Pleistocene deposit in Hutton Cave, would be specifically as indeterminable as Pomel's " Arvicola (Myolemmus) ambiguus " were it not for the fact that, under the name of " Lemtnus torquatus, var.," Sanford described and figured, at the same time and from the same cave, a fragmentary skull which evidently belongs to the same species. No more precise information as to the provenance of the skull and lower jaws than that contained in the words " from a Somersetshire cave" is given by Sanford in his paper; but the labels in his handwriting still preserved in Taunton Museum indicate that all came from Hutton Cave. The skull in question clearly belongs to the larger European Pleistocene species with cheek-teeth of the D. torquatus type, and it is convenient to treat it as though it were the type of the species, although the specific- ally indeterminable mandibular rami are the actual co-types. Characters. — Size large (dental length in adults up to 20 mm.). Cheek-teeth essentially as in D. torquatus. Skull large ; the nasals much expanded in front, their com- bined width equal to half the nasal length ; zygomatic arches very heavy ; anterior palatal foramina short and broad ; palate boldly sculptured, with the postero-lateral bridges usually in- complete ; presphenoid reduced to a slender bar ; teeth very heavy. Cheek-teeth (Fig. 71, 4) essentially as in D. torquatus, but larger and broader. In m^ and m^ the posterior walls of the liinder inner triangles are not reduced, but retain their thick enamel and primitive concave curvature; and there is in each » Hinton, Ami. Mag. N.H., [8], 6, p. 37. DICROSTONYX ICl tooth a more or less well-marked postero-internal vestigial angle ; rwg with its third or antero-external triangle less reduced than in D. henseli and frequently with a more or less well-marked minute fourth outer vestigial angle. For measurements, see table at end of volume. Remarks. — Sanford's name for this species was revived by me in 1910 after a careful comparison of the fossil material then available with the fine series of recent skulls from Discovery Bay. At that time the Discovery Bay animal was believed to be iden- tical with D. torquatus Pallas, and the comparison left and leaves no room for doubting that it is distinct from I), gulielmi. But further work has shown that the Discovery Bay Lemming is in turn specifically distinct from D. torquatus and has now to bear the name D. grcenlandicus Traill. To judge from the scanty material representing D. torquatus as now understood, the latter species makes a nearer approach to the fossil form ; but the recent material is inadequate to enable us to appreciate the specific characters of D. torquatus. The fossils are larger and have heavier teeth, and in wA and m^ the postero-external triangles are rela- tively large, that of tn^ exceeding that of m^ in size, whereas in recent D. torquatus they are smaller, subequal in size in m^ and m^ or else slightly larger in the former tooth than in the latter. On these grounds, pending the acquisition of further recent material, D. gulielmi may be maintained as a distinct species. It would at present be unwise and misleading to identify the fossils in question with what is in fact the least known of all the recent species. The following is a list of the deposits and localities in which remains have been found with short descriptions of the material upon which the records are based. In all cases, unless otherwise .stated, the material has been studied and determined by the writer. A. GREAT BRITAIN. Somersetshire. Hutlon Cave. Five mandibular rami (co-types of the species) and two frag- mentary skulls originally described by Sanford (Q.J.G.S., 26, 1870, p. 125, pi. viii, figs. 2 and 4). Re-examined by me (Fig. 71, 4 and Plate V). Taunton Museum. Somersetshire. Aveline's Hole, Burrington C'oombe. Lower jaws at first referred doubtfully to D. henseli; but much more extensive material collected later proves the species to be D. gulielmi (Hinton, Proc. Bristol Univ. Spelaeol. Soc, 1, p. 75, 1922, and 2, p. 34, 1924). Bristol Univ. Spelceological Society's Museum. Devonshire. Chudleigh Fissure. Two palatal fragments showing between them m^-m^ ; and part of a mandibular ramus. A. S. Kennard, F.G.S. 162 MICROTIN^ Herefoedshire. Cave near River Wye, Forest of Dean. Part of a skull and twelve mandibular rami ; collected, described and figured by Miss D. M. A. Bate (Geol. Mag., [4], 8, 1901, p. 104, figs. 2-6). British Mxiseum. Herefordshire. Merlin's Cave, Wye Valley, near Symond's Yat (possibly the cave previously worked by Miss Bate). Parts of twenty-two skulls and five hundred and twenty-three mandibular rami representing at least two hundred and seventy individuals (Hinton, Proc. Bristol Univ. Spelseol. Soc, 2, p. 157, 1925). Many more not yet reported upon were obtained last autumn. Bristol University Speloeological Society's Museiim. Wiltshire. Brickearth of Fisherton, near Salisbvry. Remains originally determined by Blackmore and Alston as Myodes torqnalus (P.Z.S., 1874, p. 470). Dr. Blackmore has kindly lent two mandibular rami from this deposit; they are difficult to determine, but I think both are referable to D. guUelmi. Blackmore Museum, Salisbury. Kent. Erith. Middle Terrace brickearth of the Thames. Part of a skull and mandible collected by Mr. F. C. J. Spurrell ; described and figured, under the name Myodes torquatus, by Mr. E. T. Newton (Geol. Mag., [3], 7, 1890, p. 454, figs. 3-6). Re-examined and determined as D. gvlielmi by the writer; this record supports the opinion that the Fisherton remains are to be referred to D. gulielmi, because in other respects the fauna of the Fisherton brickearth is closely similar to that obtained at Erith. Museum of Practical Geology. Derbyshire. Langwith Cave, near Mansfield. Part of a skull, several mandibular rami and detached maxillary molars; collected by the Rev. E. H. Mullins (Derbyshire Archseol. and N.H. Soc. Journ., 1913, p. 15 of separate). Lancashire. Dog Holes, Warton Crag. Part of a right maxilla with m^ in place; a detached right m^ and six mandibular rami in British Museum presented by the collector, Mr. J. Wilfrid Jackson. Other material in Mr. Jackson's collection at Manchester not seen by me (see Lancashire Nat., 1909, p. 227). British Museum ; Manchester. B. IRELAND. Co. Sligo. Kesh Caves. Co. Cork. Edenvale Caves. Numerous mandibular rami, a few fragmentary maxiUae and detached upper molars. Dublin Miiseiim. C. CONTINENTAL EUROPE. Southern France. Neschers. A fragmentary palate with m^ in place and some other remains probably referable to this species. British Museum. DICROSTONYX 163 8. fDicrostonyx henseli Hinton. 1855. Misothermus torqiiatus HenscI, Zeitschr. deutsch geol. Gesellsch., 7, p. 493, taf. XXV, figs. 12 and 13. Not of Pallas, 1779. 1872. Myodes torquatus Forsyth Major, Atti Soc. Ital. Sci. Nat. Milano, 16, p. 123; Newton, Q.J.G.S., 50, 1894, p. 196. 1910. IHrrostonyx henseli Hinton, Ann. Mag. Nat. Hist., [8], 6, p. 37. Type.—BM., No. M.11,803, Geol. Dept. ; Abbott Collection; presented by Sir H. H. Howorth, K.C.I.E., F.R.S. Type locality and Horizon. — Ightham, Kent. Late Pleisto- cene; from a deposit filling fissures in the "Kentish Rag." Characters. — Size rather small (dental length in adults up to 19 mm., condylo-basal length not exceeding 29 mm.). Cheek- teeth relatively broad, but essentially like those of D. hudso)nus in pattern. Skull (Fig. 70) in general form most nearly resembling that of D. hudsonius but smaller; its dorsal contour gently con- vex from before backwards, as in D. hudsonius, but with smaller and less expanded nasals, and relatively shorter diastema ; palate feebly sculptured, with complete postero-lateral bridges ; presphenoid reduced to a slender rod, in correlation with the relatively broader and heavier cheek-teeth ; auditory bullae very small, egg-shaped and not inflated anteriorly, their general form much as in D. hudsonius but their position a little different, lying with their long axes more convergent anteriorly and less nearly parallel with the long axis of the skull. Cheek-teeth (Figs. 71, 8 ; 72, 7) relatively broad, but essentially like those of D. hudsonius in pattern. In m^ and ?/i- the posterior walls of the hinder inner triangles are reduced, losing their primi- tive concave curvature and tending to lose their thick enamel; a postero-internal vestigial angle never developed ; m^ with a small antero-internal vestigial angle, but no corresponding outer vestige, the third outer triangle more or less reduced. Other parts of skeleton not essentially different from those of the hving members of the genus. For measurements, see table at end of volume. The following is a list of deposits and localities in which remains of D. henseli have been found. In all cases, unless otherwise stated, the material has been determined by the writer. A. GREAT BRITAIN. Kent. Ightham, near Sevenoaks; fissure deposit. A nearly complete skull (the type); an incomplete skull and several mandibular rami. British Museum and Museum of Practical Geology {e.v W. J. Lewis Abbott Collection); and in collection of Dr. Frank Comer. Middlesex. Lea Valley ; Third Terrace deposits at Bonder's End and Angel Road. Abundant remains, including a nearly complete skeleton, and 164 MICEOTINJE quantities of the dung preserved in peaty silt (Hinton, Q.J.G.S., 68, 1912, p. 249; Warren, ibid., p. 213, and 71, p. 175, 1916). Derbyshire. Langwifh Cave, near Mansfield. Two or three examples of m^ and ni- were referred by me to D. henseli (Mullins, Derbyshire Archseol. and N.H. Soc. Journ., 1913, p. 15 of reprint). Herefordshire. Merlin's Cave, Wye Valley, near Symond's Yat. Four skull fragments with m^ and inr in place seem to be refer- able to this species and not to D. gulielmi, which occurs so abundantly in this cave (Hinton, Proc. Bristol. Univ. Spelaeol. Soc, 2, p. 157, 1925). Bristol University Spelceological Society's Museum. A fragmentary right mandibular ramus from a Pleistocene deposit at Corstorphine, west of Edinburgh, determined by E. T. Newton as Dicrostonyx sp., has been recorded by W. Evans (Scott. Nat., 1913, p. 97). B. IRELAND. Co. Clare. Doneraile Cave. Several fragmentary palates and numerous mandibular rami. Dublin Museum. C. CHANNEL ISLANDS. Jersey. La Cotte de St. Brelade. A palate and a number of mandibular rami indicating at least fifteen individuals (Hinton, Soci6te Jersiaise Bull. Ann., 43, 1918, p. 355). D. CONTINENTAL EUROPE. Northern France. Manbeuge ; fissure deposit. Fragmentary skulls and lower jaws described and figured, under the name Myodes torquatus, by G. Dubois (Ann. Soc. Geolog. du Nord, 44, p. 69, 1919). The figures of the upper molars indicate that the species represented by these remains is D. henseli. Cambrai ; " dans un sable situe a la base de Vergeron." Remains described by G. Dubois (Ann. Soc. Geolog. du Nord, 44, p. 95, 1919) ; probably referable to D. henseli. South Germany. Wurttemberg, Kleine Scheuer. Numerous palates and several mandibular rami; lent by Dr. Frank Corner. Saxony, Quedlinburg. The skull and other remains described by Hensel (Zeitschr. deutsch. geol. Gesellsch., 7, p. 493). Remains of Dicrostonyx have been recorded chiefly by Nehring and Woldrich from a great many late Pleistocene deposits on the Continent, but in the absence of material I am unable to determine the species. A nearly perfect skull from Eppelsheira, near Darmstadt, mentioned by Blackmore and Alston (P.Z.S., 1874, p. 470) as being in the British Museum was accidentally destroyed many years ago; the mandibute which remain are insufficient for specific determination. 1 SYNAPTOMYS 165 Genus : 2. SYNAPTOMYS Baird. 1857. Syiiaplomys Baird, Mamm. N. America, p. 558, Genotype. — Synaptomys cooperi Baird. Range. — ^North America, from the northern edge of the Lower Austral zone in Virginia (Dismal Swamp) and Kansas (Woodson County) northwards to Alaska, Mackenzie, and Labrador. Characters. — General external form nearly as in normal voles. Fur soft and moderately long. Colour dull ; bi'ownish above, paler below. External ears well developed, slightly evident above the fur; in form like those of Evotomys. Hands and feet of normal size and form ; thumb small, jjrovided with a large flattened nail; five palmar and six plantar tubercles; palms clothed between wrist and jaads with long white hairs which conceal the jjollical tubercle; soles clad with shorter and thinner haii-s between the heel and pads. Tail terete, slightly longer than the hind-foot; rather thinly clothed with long stiff hairs which do not com2:)letely conceal the scaly annulations. Mammae, 2 — 2 = 8 or 1 — 2 = 6. Skull relatively narrower, more lightly built, and in general less modified than in Leiiimus. Rostrum noticeably deflected, short and broad. Nasals ending anteriorly slightly but dis- tinctly in advance of the strongly curved " opisthodont " upper incisors ; terminating posteriorly in line with the premaxillaries and slightly in front of the orbits. Zygomatic arches lighter and less widely S2)reading than in Leminus, though more expanded than is usual in voles ; the greatest zygomatic breadth falls on the maxillary or fore-parts of the arches and is equal to 61-65% of the condylo-basal length, instead of to about 68% as in Lemmus or to 55-60% as in most voles. Jugals slender, with slightly convex upper borders, their outer surfaces forming nearly vertical planes instead of being abruptly convergent dorsally as in Lemmus. Braincase longer and narrower than in Lemmus, the post-glenoid region less noticeably shortened. Temjioral ridges fusing ^ in old age to form a median inter- orbital crest; their posterior course over the braincase essentially as in Le>nmus, though they are perhaps slightly less salient than in equal-aged skulls of L. leminus. Squamosals very large, essentially as in Lemmus, though relatively a little more widely separated in front than in the latter genus, with long post- orbital crests, which bound the square shoulders of the braincase, and forming the entire floors of the posterior or supratym2)anic * In two only (.S'. wrangeli, Metlakatla, British Columbia, B.M., Nos. 11.10.2.1 and 13.10.1.1) of the small series of skulls of Synaptomys before me, are the ridges actually fused into an interorbital crest; but in several of the other specimens the ridges are closely appro.ximated. Jlost probably the ridr;es fuse regularly in the older adults of all species of Synnptoinyn; but fully mature skulls of Microtinae are naturally comparatively rare. 166 MICROTINiE portions of the temporal fossae ; supratympanic fenestrse relatively large ; owing to the slighter lateral saliency of the lambdoid crest, the supratympanic portions of the temporal fossae are less extensive transversely than in Leinmus. Interparietal rather longer in relation to its width, truncated laterally (although not crossed by the temporal ridges), and usually with a boldly convex posterior border. In the ventral view the great width and shortness of the rostrum are consjiicuous. Antero-palatal foramina rather short but unusually broad. Cheek-teeth narrower in proportion to their length, the tooth rows, about equal to the diastema in length, diverging posteriorly, though not so rapidly as in Leninius. Palate posteriorly nearly as in Microtus ; the posterior median sloping septum longer and better developed than in Leminus, the inner or mesial borders of the post-palatal pits not extending so far forwards but bending inwards to make a junction with the sides of the median septum whereby an appearance, closely similar to that seen in many species of Microtus, is imparted to the hinder edge of the palate. Ptery- goid fossae slightly longer but a little shallower than in Lemmus, their floors being scarcely dorsal to the ventral surface of the basisphenoid. Choanse, presphenoid, basisphenoid, and fore-part of basioccipital all considerably narrower than in Leuivtus. Auditory bullae relatively large and globular, more closely ap- proximated to the middle line below, their inner sides being conspicuously more inflated than in Lemmus ; in each bulla the external meatus is very shortly tubular; the walls are densely spongy, the sponge being connected with the surface of the petrous portion by a wide mesh work of strong bony trabeculae; canaliculus tympanicus ossified until it reaches the staj^es ; mastoid portion and tegmen tympani less inflated than in Le>iimus. Mandible essentially as in Lemmus, but condylar jjrocess in each ramus shorter and the groove between the cheek-teeth and the ascending ramus more evenly and smoothly continuous with the concave inner surface of the angular process behind. Dentition,. — Incisors like those of Lemmus as regards their courses and lengths. Ujiper incisors strongly curved, " opistho- dont," each with a longitudinal groove near the outer edge of its anterior surface ; cutting edges of upjjer incisors as in Lem- mus, when uninjured, the soft dentine worn back and hollowed considerably, the hard and resistant anterior enamel jilate becoming very salient and forming a tubular or gouge-like termination. Cheek-teeth rootless ; infolds filled with cement ; enamel well diiJerentiated into thick and thin jiortions forming respectively the concave and the convex sides of the salient angles. Enamel pattern essentially as in Lemmus ; but in some species jiersistent though quite vestigial traces of the median row of tubercles occur frequently (Fig. 73). In m^ the second and third transverse SYNAPTOMYS 167 f^ig.la '"'9 23 Fig. 3a. Hg lb. fig.2b. Fig 3b Fig. 73. — Cheek-teeth of Synaptomys. Crown view3 : a. upper molars ; 6. lower molars : — 1. S. (Symtptomys) fatuuti. Ontario. 2. S. (Miclomys) .sp. Metlakatla, British Columbia. 3. S. (Mictomys) borcalU. Type. 168 MICROTINiE loops are separated by the very deep second outer fold, the first inner infold being but slightly develo])ed. In the lower molars the outer infolds may be deep enough to close triangles (subgenus Synaptomys, Fig. 73, 16) ; or so weakly developed that the teeth consist of transverse loops and have crenulate instead of serrate outer borders (subgenus Mictomys, Fig. 73, 2&, 3&). Subgenera and species. — Eleven species, comprising thirteen forms, are at present included in the genus ; but future work will probably reduce the number of species and increase the number of geographical races. The known forms may be arranged in two subgenera, viz., Synaptomys Baird and Mictomys True, distinguished as follows : — • Mammae, 1 — 2 = 6. Lower cheek-teeth with closed triangles. Rostral part of skull very stout; palate without long posterior sj3ine. Subgenus tSynaptomys Baird. Mammae, 2—2 = 8. Lower cheek-teeth without closed triangles. Rostral part of skull slender ; palate with long spinous process behind. Subgenus Mictomys True. Subgenus : SYNAPTOMYS Baird 1857. Synaptomys Baird, Mamm. N. America, p. 558 (genus). 1896. Synaptomys Miller, N. Amer. Fauna, No. 12, p. 34 (subgenus of Synaptomys); Merriam, Proc. Biol. See. Washington, 10, 1906, p. 57. Genotype. — Synapto)nys cooperi Baird. Range. — -Eastern North America, from Virginia and Kansas northwards through the eastern United States into eastern Canada and westwards to Minnesota. Characters. — Externally as described under the genus. Mammae, 1 — 2 = 6. Skull with remarkably heavy rostrum ; palate nearly as in Microius, without long posterior spine. Incisors very large and broad ; grooves of upjjer incisors usually well defined and externally placed. Mandibular cheek-teeth with deep outer infolds and closed triangles substantially as in Lemmus. Geographical differentiation. — ^Four forms referred to three sjDccies are at present recognized ; but according to Rhoads (Proc. Acad. Nat. Sci. Philadelphia, 1897, pp. 305-307) all should be regarded as subspecies of S. cooperi. Only one of these forms is well represented in the material before me, and I can therefore come to no decision ; but after carefully studying the literature and my material I am inclined to think that Rhoads is right. SYNAPTOMYS 169 1. Synaptomys (Synaptomys) cooperi Baird. 1857. Synaptomys cooperi Baird, Mamm. N. America, p. 558. 1893. Synaplomi/fi slonel Rhoads, Amer. Nat., 27, p. 53; described from May's Landing, Atlantic County, New Jersey; type in collection of S. N. llhoads. 1894. Si/naptomi/s cooperi Bangs, Proc. Biol. Soc. Washington, 9, p. 99; Miller, " List," 1924, p. 394. Co-types. — U.S. National Museum. Type locality. — Unknown ; j^robably northern New Jersey or southern New York. Range. — Eastern United States from Minnesota eastwards to eastern Massachusetts ; south to Iowa, Indiana, and Mary- land, and in the mountains to North Carolina and Tennessee. In the Boreal and parts of the Transition Zones, finding,' a boreal atmosj^here south of the Boreal Zone in cold sphagnum swamps. Characters. — Distinguished from *S'. helaletes by smaller feet and lighter skull. Hind-foot 18 mm. Colour of ujiper parts grizzled grey and yellowish brown abundantly mLxed with black-tipped hairs ; under j^arts dirty white darkened by the hair-bases. Tail bicoloured, brownish above, whitish below. Skull and teeth relatively small and weak comjjared with those of