/^BERKELEY* LIBRARY I UNIVERSITY OF \CAUFORNIA EARTH SCIENCES LIBRARY Cambridge (geographical GENERAL EDITOR : F. H. H. GUILLEMARD, M.D., FORMERLY LECTURER IN GEOGRAPHY IN THE UNIVERSITY OF CAMBRIDGE. A GEOGRAPHICAL HISTORY OF MAMMALS. ilontion: C. J. CLAY AND SONS, CAMBRIDGE UNIVERSITY PRESS WAREHOUSE, AVE MARIA LANE. ffilaggofo: 263, ARGYLE STREET. : F. A. BROCKHAUS. lorfe: MACMILLAN AND CO. THE WORLD ON MERCATOR'S PROJECTION SHEWING THE Cambridge TJniversi UVIMALIAN GEOGRAPHICAL REALMS & REGIONS. 90- ins- i»o' London^: Stamfords Geog*' Estab*. A GEOGRAPHICAL HISTORY OF MAMMALS, BY R. LYDEKKER, B.A., F.R.S., V.P.G.S. ETC. CAMBRIDGE: AT THE UNIVERSITY PRESS. 1896 [All Rig/its reserved.} PALEONTOLOGY LIBRARY Gift of C. A. Kofoid (JEamfartoge : PRINTED BY J. AND C. F. CLAY, AT THE UNIVERSITY PRESS. L3 PREFACE. SCIENCES LIBRARY SINCE the publication of Dr Wallace's book on the geo- graphical distribution of animals in general, the only works which have appeared relating to Mammals from the same point of view are the small volume by Mr F. E. Beddard and the series of papers by Mr W. L. Sclater, referred to in the Appendix. Both the latter admittedly take but little account of fossil forms ; and there is accordingly ample room for a work which should collect and arrange the information on this subject, and indicate the deductions which may be drawn therefrom. This task has been attempted in the present volume. The subject is, however, so vast, and the information relating to it scattered through so many publications, that it is probable many points of interest or importance have not been noticed. From the mode of arrangement of the work, a considerable amount of repetition was inevitable. The long time that the volume was in the press will account for its containing no allusion to certain papers which have ap- peared recently. It may therefore be mentioned here that recent discoveries have shown a slight intermixture of northern types in the Palaeozoic flora of Argentina, so that the isolation of the northern and southern floras is .not so nearly complete as was supposed. A paper just published by M. Boule1 indi- 1 Comptes RenduS) vol. cxxii (1896). VI PREFACE. cates that the relationship between Cadurcotherium and the Astrapotheria is closer than suggested on page 82. Hence there is further evidence that South America received its Tertiary ungulates by way of Africa. The extinct Patagonian birds Phororhachis and Brontornis appear, according to Mr C. W. Andrews, to be nearly allied to the existing South American seriema (Cariama), and since Filholornis, of the French Phos- phorites, has been shown by Prof. A. Milne-Edwards to be related to the hoatzin (Opisthocomus), it would seem that all these peculiar South American types of birds have a history somewhat similar to that of the extinct ungulates of the same region. R. LYDEKKER. HARPENDEN, June ist, 1896. CONTENTS. CHAPTER I. INTRODUCTORY. Distributional Area and Station — Influence of Temperature — Humidity — Other Factors in Distribution — Importance of Palaeontology — Inequality in the Ages of different Groups of Animals — Different Groups have different Geographical Distribution — Importance of Mammals in Geographical Distribution — Classification of Mammals — Barriers to Dispersal of Mam- mals— Influence of Man on Distribution — Extinction of the larger Plistocene Mammals — Distributional Areas of Genera and Species — Centres of Evolution — Permanency of Continents and Ocean-Basins — Zoological Realms and Regions ....... i CHAPTER II. THE NOTOG^EIC REALM. Definition and Characters of the Realm — Australian Region — Monotremes — Marsupials — Rodents — Carnivores — Ungulates — Bats — List of Australian and Papuan Genera — Polynesian Region — Hawaiian Region — Austro- Malayan Region — Palseontological History of Marsupials — How Australia received its Fauna . . . . . . . . .28 CHAPTER III. THE NEOG^EIC REALM. Extent and Characters — Mammaliferous Deposits — Monkeys — Bats — Insecti- vora — Carnivores — Ungulates — Horses — Litopterna — Astrapotheria — Toxodonts — Pyrotheria — Proboscideans — Rodents — Edentates — Arma- dillos and Glyptodonts — Sloths — Anteaters — Ground-sloths — Marsupials — Cetaceans — Early Distinction of the Neogseic Fauna — Early Separation of N. and S. America — Incursion of Northern Mammals — Distinctness of the existing Fauna — Origin of the Santa Cruz Fauna — Antarctica and the South American element in the Ethiopian Fauna — Conclusion — Sub- regions 64 Vlll CONTENTS. CHAPTER IV. THE ARCTOG^EIC REALM. Features of the Arctogaeic Fauna — Community of earliest Fauna — Evidence of Secondary Reptiles — Puerco Fauna — Lemuroids — Insectivora —Carnivores — Rodents — Ungulates — Summary of the Characteristics of the Mamma- lian Fauna of Arctogaea . . . . . . . . 1 44 CHAPTER V. EASTERN ARCTOG^EA. Mammalian Groups peculiar to Eastern Arctogsea — Tertiary Mammalian Fauna of Eastern Arctogsea — Oligocene Fauna — Miocene Fauna — Older Pliocene Fauna — Pikermi and allied Faunas — Sivvalik Fauna — Higher Pliocene Faunas . . . . . . . . . • 1/9 CHAPTER VI. THE MALAGASY REGION. Mammalian Fauna — Relations of Madagascar to the Mainland . . 213 CHAPTER VII. THE ETHIOPIAN REGION. Characteristics of the Mammalian Fauna — Birds — Past History of Ethiopia — Sub-regions . . . . . . . . . . .227 CHAPTER VIII. THE ORIENTAL REGION. Sub-regions — Characteristics of the Mammalian Fauna — Past History of the Region — Malayan Sub-region — Nicobars, Mentawi, and Christmas Islands — Philippine Sub-region ......... 264 CONTENTS. IX CHAPTER IX. THE HOLARCTIC REGION. Characteristics of the Mammalian Fauna — Mammals of the Eastern Holarctic Region — Plistocene Fauna of the Holarctic Region — Geographical Changes since the Plistocene — Western Division of the Region — Arctic Sub-region — European Sub-region — Central Asian Sub-region — Tibetan Sub-region — Mantchurian Sub-region — Mediterranean Sub-region — Kashmir — Ca- nadian Sub-region — Transition Zone . . . . . . 308 CHAPTER X. THE SONORAN REGION. Limits — Characteristics of Mammalian Fauna — Extinct Groups of Mammals characteristic of Western Arctogrea — Distinctness of the Region — Dual origin of Groups . . .. . . .-..'.'' V '•' '. 363 APPENDIX. LIST OF WORKS AND PAPERS .- &*&, . . ..:..-, fj . 383 INDEX . . . . V . 386 MAP SHEWING DISTRIBUTION OF ANIMALS . . . , to face Title LIST OF ILLUSTRATIONS. PAGE Fig. i. The Duckbill {Ornithorhyncus anatinus] . . . .32 „ 2. Skull of Rat-Kangaroo 35 ,, 3. Skull of Extinct Phalanger ( Thylacoleo carnifex) 37 „ 4. Banded Anteater (Myrmecobius fasciatus] .... 39 ,, 5. Fore Part of Skull of Babirusa (Babirusa alfums} . . -47 ,, 6. Lower Jaw of Triconodon, or Allied Form . . . .52 ,, 7. Lower Jaw of Phascolotheriiim . . . . . -53 ,, 8. Lower Jaw of Amphilestes . . . . . . -53 ,, 9. Lower Jaw of Homunculus ....... 69 ,, 10. Skeleton of Macrauchenia patachonica ..... 76 ,, n. Skeleton of Hind Foot of Rhinoceros 78 ,, 12. Carpus and Metacarpus of Hyrax and Elephant . . -78 ,, 13. Upper Molar of Macraiichenia .80 ,, 14. Upper Molar of Oxyodontotherium ...... 80 ,, 15. Upper Molar of Astrapotherium . . . . . . 81 „ 16. Skull of Nesodon 83 ,, 17. Palate of Typotherium ........ 84 „ 1 8. External Skeleton of Armadillo ( Tatusia giganted] . . . 93 „ 19. Internal Skeleton of a Glyptodont ...... 95 ,, 20. External Skeleton of Glyptodon clavipes . . . . -97 ,, 21. External Skeleton of Panochthus tuberculatus . . . -99 ,, 22. Tamandua Anteater 102 ,, 23. Skull of Mylodon ......... 104 ,, 24. Skeleton of Scelidothcrium leptocephalum . . . .106 „ 25. Lower Jaw of Prothylacinus ....... 108 ,, 26. Lower Jaw of Acdestis oweni . . . . . . .no ,, 27. Lower Jaw of Abderites . . . . . • • . rii ,, 28. Lower Jaw of Plagiaulax 149 ,, 29. Skull of Tritylodon 150 ,, 30. Molar of Tritylodon • • -150 „ 31. Upper Cheek-Teeth of Plesiadapis 155 ,, 32. Upper Cheek -Teeth of Microcharus 156 LIST OF ILLUSTRATIONS. xi PAGE Fig- 33- Upper Molars of A rctocyon 159 ,, 34. Upper Molar of Ancodus . . . • . . . . 161 ,, 35. Skeleton of Elotherinm 162 „ 36. Upper Cheek-Teeth of Palaotheriiim 166 „ 37. Upper Cheek-Teeth of A nchitherium , . . . . 167 ,, 38. Upper Molar of Horse 167 „ 39. Skull of Hipparion . . . >. , . . ... .168 ,, 40. Upper Molar of Rhinoceros . . ..' , i, . •„ »•;- . ,*.,.. 169 ,, 41. Skeleton of Titanotherium . . . • » .-• k, . 171 ,, 42. Upper Molar of Mastodon . . .. : .;•...: • >.•"••. - . 173 ,, 43. Upper Cheek-Teeth of Coryphodon . . •<• .,...»?!.;* . 174 ,, 44. Upper Cheek-Teeth of Anoplotherium .. \ ,'. , . , ,. ... ; . 185 ,, 45. Skull of PalcBorias . •+< -•; *,< '.;/•»*- .1. . ». .. •'. .. . 199 , , 46. Upper Molar of Merycopotanius . . :. 1 • < c, . . 203 ,, 47. Restoration of Sivatheriutn . . . . -. . »v •/»... . 205 ,, 48. Skull of Lemur . , »' . . . . .; • '» ; . 217 ,, 49. Ring-tailed Lemur (Lemur catta) * . ; «'\ '.• '{. • rt&l . . 218 ,, 50. Skull of Aye-aye (Chiromys madagascariensis} .«,,.. . 219 ,, 51. The Fossa (Cryptoproctaferooc] . ... *,;,., • •• • 221 ,, 52. African Jumping-Shrew (Macroscelides tetradactylus} . . 232 >» 53- West African Potamogale (Potamogale velox] ;.-. ... . 233 „ 54. Cape Hunting-Dog (Lycaon pictus] . .i ....... . 236 ,, 55. Fulgent African Flying Squirrel {Anomahirus fiilgens) . . 237 ,, 56. Head of African Wart- Hog (Phacochcerus cethiopicus) . . 241 ,, 57. Water-Chevrotain (Dorcatherium aquaticuni) .... 243 ,, 58. Head of Gemsbok (Oryx gazelld) . . . . . . 246 ,, 59. Cape Hy rax (Procavia capensis) .... . . 248 „ 60. Slow Loris (Nycticebus tardigradus} ..... 269 ,, 6r. Tree-Shrew (Tupaia tana) 271 ,, 62. Indian Sloth-Bear (Melursus ursinus) 275 ,, 63. Indian Ratel (Mellivora ratel) . 276 ,, 64. Japanese Serow (Nernorh&dus crispus] 279 ,, 65. White-bellied Pangolin (Manis trictispis) .... 283 ,, 66. Russian Desman (Myogale moschatd] . . . . 320 ,, 67. Head of Spanish Ibex (Capra pyrenaica} 323 ,, 68. Musk-Deer (Moschus moschiferus] . . . . . 325 ,, 69. Rocky Mountain Goat (Haploceros montanus] .... 343 ,, 70. Musk-Ox (Ovibos moschatus) . . . . . i . 346 , , 71. Musquash (Fiber zibethicus) . 362 ,, 72. Face of Geomys bursarius . . . . * . . . . 366 ,, 73. Face of Thomomys talpoides . . . . ... . 366 „ 74. Foot of Geomys 377 ,, 75. Head of Mule-Deer (Cariacus macrotis) 368 ,, 76. Head of Prong-buck (Antilocapra americana) .... 369 xii LIST OF ILLUSTRATIONS. PAGE Fig. 77. Skeleton of Patriofelis ferox ..... • 372 ,. 78. Skeleton of Agriochcerus latifrons .... • 374 »• 79- Hind-foot of Agriochcerus ..... • • 376 ,. 80. Skull of Protoceras ...... • • 376 ,, 81. Molar of Palceosyops • 377 „ 82. Extremity of Skull of Uintatherium • • 378 Figs, i, 4, 22, 28, 32, 48, 49, 50, 52, 53, 54, 55, 57, 59, 60, 63, 65, 66, 68, 70, 71 are from The Study of Mammals > Living and Extinct, by Flower and Lydekker. Figs. 13, 14, 15, 16, 17, 18, 19, 20, 21, 23, 24 are taken from the author's work on Argentine Fossil Mammals, published in the An. Mus. La Plata. Figs. 9, 25, 26, 27 are from Senor Ameghino. Fig. 67 is from a photo- graph by Mr Abel Chapman. Fig. 82 is from Prof. Cope, fig. 81 from Prof. Earle, fig. 40 from Prof. Boyd Dawkins, fig. 30 from Prof. Fraas, and figs. 7, 8 from Prof. Goodyear. Fig. 46 is taken from a plate in Hutchinson's Extinct Monsters ; 58 from a photograph by Mr Eccles, of Reading, and 75 from one in the possession of Mr J. E. Harting. Figs. 6, 35, 41, 43 are from Prof. Marsh; and n, 12, 31, 33, 37, 78, 79, 80 from Prof. H. F. Osborn. For the blocks of figs. 72, 73, 74, the author is indebted to Dr C. H. Merriam. Fig. 45 is from Nicholson and Lydekker's Manual of Paleontology ; fig. 29 from Owen; fig. 64 from Dr Sclater in the Proc. Zool. Soc.; fig. 77 from Dr Wortman, and 39 from Prof, von Zittel. A GEOGRAPHICAL HISTORY OF MAMMALS. CHAPTER I. INTRODUCTORY. Distributional Area and Station — Influence of Temperature — Humidity — Other Factors in Distribution — Importance of Palaeontology — Inequality in the Ages of different Groups of Animals — Different Groups have different Geographical Distribution — Importance of Mammals in Geographical Distribution — Classification of Mammals — Barriers to Dispersal of Mam- mals— Influence of Man on Distribution — Extinction of the larger Plistocene Mammals — Distributional Areas of Genera and Species — Centres of Evolution— Permanency of Continents and Ocean-Basins — Zoological Realms and Regions. THAT there are differences in the animals and plants of different districts and different countries is a fact apparent to every person who has travelled at all; while to those who have travelled ex- tensively it will further be evident that the amount of this difference is by no means correlated with the distance of one country from another, the fauna of Japan, for instance, being much more like that of England and France than is the fauna of Eastern Africa to that of the adjacent island of Madagascar. Unfortunately, among persons who are not conversant with the principles of zoological science, the distinction between the faunas of different countries has been much obscured by the practice common to almost all the old voyagers and colonists of bestowing upon the animals of new countries the names of such Old World creatures as they appeared most nearly to resemble. The puma of America was, for instance, called the lion, and the jaguar of the same country, the tiger; while the koala of Australia was christened the native bear, and its marsupial allies the dasyures are still commonly spoken of as native cats. To students of the science of L. I INTRODUCTORY. [CHAP. Geographical Distribution it is, therefore, essential to discard such misleading popular titles, and to speak of animals by their correct names. Apart from the specific or generic distinctions between the animals of one country and another, the observer Distributional Area and sta- will not fail to notice more or less well-marked differences between those inhabiting different dis- tricts of a single country ; such differences being most intensified when a country presents great variation in its physical features. An excellent instance of this is afforded by South America, where there are the open grassy plains of the Argentine, the dense tropi- cal forests of Paraguay and Brazil, and the snow-clad heights of the Andes. In the former tract the traveller will meet with the peculiar rodents known as viscachas, the Patagonian cavy, a species of deer, numerous armadillos, and the rhea (miscalled the American ostrich). In the Brazilian forests, on the other hand, he will find monkeys, marmosets, tapirs, tree-porcupines, sloths, and anteaters, together with certain armadillos which are for the most part speci- fically or generically distinct from those of the pampas. If, on the other hand, he ascend high on the Andes, he will leave behind the animals of the forest, to be confronted with chinchillas, guanacos and vicunas. Different, however, as are the animals of these various districts, yet an acquaintance with their zoological affinities will prove that many of them belong to closely allied groups, some of which are met with in no other parts of the world. This will serve to show that they belong to what is known as one zoological province or region, and that the differences between the faunas of different districts of that province are due to the physical variations between its component districts. Perhaps this point may be still better illustrated by the cases where the same species of animal is restricted to different districts of one country or continent. For instance, the common squirrel is only found in the wooded districts of Europe, and is entirely absent from open plains. The chamois, again, is only met with in the isolated mountain ranges of the Pyrenees, the Alps, and the Caucasus ; while the Siberian ibex of the Altai reappears in Tibet and the Himalaya, but is wanting in the intervening tracts. In these instances Europe would be spoken of as the distributional I.] INFLUENCE OF TEMPERATURE. 3 area of the squirrel and the chamois, and Central Asia as that of the Siberian ibex ; but the particular districts suited to the exist- ence of each would be termed its station. And here it may be mentioned that whereas the distributional area of a species is generally continuous, its various stations may be partially or com- pletely isolated, as in the instances of the aforesaid mountain animals, which cannot live on the plains below. Station is thus seen to be very intimately connected with temperature; and of this a very striking example may be found among the mammals of South America. As already mentioned, the llama-like animals respec- tively known as vicunas and guanacos are met with in company on the highlands of the Cordillera in Peru and Ecuador, but as we go further south the latter are found on the plains of southern Argentina and Patagonia, as well as in the island of Tierra-del- Fuego, at the sea-level. Here, then, there is a clear proof of the intimate connection existing between temperature and station; the guanaco, being an animal which can live only in cold or temperate climates, finds suitable conditions for its existence in tropical latitudes solely at a height of many thousands of feet, although further south it is able to thrive at the sea-level. This being so, it is obvious that temperature must likewise exert a very considerable influence on the whole distributional area of many animals. Of this, the most marked instance is found in the fauna of the Arctic regions, which forms a circumpolar zone of animals more or less markedly distinct from those dwelling further south. And if the whole land-area of the world were con- nected, and not broken up by mountain-chains, its faunas might probably be divided into zones or belts, whose limits would mainly depend upon temperature. In this connection it may be mentioned that the instance of the range of the guanaco is of considerable importance in regard to a decided difference between the Old and New Worlds in respect to the influence of mountains on the present distribution of the animals of the two areas. In the Old World the chief mountain- ranges, such as the Pyrenees, Alps, Carpathians, Caucasus, Hindu- Koh, Himalaya, Thian Shan, and Altai, run in a more or less decided east-and-west direction; whereas in America, and more I — 2 INTRODUCTORY. [CHAP. especially in the southern half of that continent, they have a north- and- south trend. Consequently, whereas in the former area the mountain-ranges have acted as barriers to the dispersal of animal life, there has been no such obstacle to diffusion in America, and animals have been able to distribute themselves according to tem- perature-conditions. It is in consequence of this physical feature that a single species of cold-loving animal like the guanaco can range at the present day from the equator to latitude 55° south, whereas in the Old World the common ibex is now restricted to the isolated mountain-ranges of Europe, and the Siberian ibex is confined to the systems of the Himalaya and the Altai. Such isolated stations could, of course, only have been reached during a period when the general temperature of the northern hemisphere was much colder than it is at present, and the animals were thus enabled to cross the lowlands from one chain to the other ; and that such a cold period formerly existed, there is abundant evidence in the traces of an extensive glaciation found over a large portion of Europe and Asia. Although, therefore, strictly speaking, tem- perature has really had as much effect in the distribution of animals in the Eastern Hemisphere as in the Western, yet, since the glacial epoch, its influences have been considerably masked by the trend of the chief mountain-ranges, and likewise by the greater isolation of the different countries of the former area as compared with the latter. These essential differences thus render it impossible to mark out the Old World in the zones of animal distribution which have been attempted for North America ; and they will likewise serve largely to explain the divergence of views on this point which may be noticed between the writings of Drs Wallace and Merriam. The latter writer1, whose conclusions are mainly based on the evidence of North American animals and plants, is of opinion that in the northern hemisphere " animals and plants are restricted in southward distribution by the mean temperature of a brief period covering the hottest part of the year"; and it is added that in certain districts the mingling of essentially northern types with those characteristic of a more southerly zone is due to the mean temperature of the hottest part of the year being sufficiently low 1 Appendix, No. 20. I.] FACTORS IN DISTRIBUTION. 5 for the existence of the former, while the total quantity of heat suffices for the latter. In other words, there is a low summer temperature combined with a high total sum of heat. Among the secondary causes affecting distribution, humidity, according to the same observer, may occupy the first place. "Humidity," he writes, " governs details of distribution of numerous species of plants, reptiles and birds, and of a few species of mammals, within the several temperature-zones. . . . Humidity and other secondary causes determine the presence or absence of particular species in particular localities within their appropriate zones, but temperature predetermines the possibili- ties of distribution ; it fixes the limits beyond which species cannot pass ; it defines broad trans-continental belts within which certain forms may thrive if other conditions permit, but outside of which they cannot exist, be the other conditions never so favourable." Important as the influence of temperature and, in a smaller degree, that of humidity, has undoubtedly been in other Fac determining the distributional limits of the species tors in Distri- f. , , , . , , . . bution. or genera ot animals and plants now inhabiting particular countries, and large as has been the part played by the glacial epoch in producing the present condition of things, it is evident that temperature has been by no means the only, even if it be the chief factor in distribution. In the first place there are several species, more especially among the carnivorous mammals, which seem quite independent of both station and temperature, the New World puma ranging from Patagonia to Canada, while the tiger inhabits alike the burning jungles of India and Burma, and the Arctic tundras of Siberia. Striking as such cases are, they are, however, to be regarded merely as examples of the individual adaptability of certain species, which, like the carnivores named, are able to obtain suitable food in any part of the world, and they do not throw any discredit on the power of temperature as a controlling factor in animal distribution generally. Of the utmost importance in this respect are the changes which the surface of the globe itself has undergone in past epochs, whereby continents that are now more or less completely sundered from one another were formerly connected, while what INTRODUCTORY. [CHAP. are now islands were once parts of continents, and vice versa. Such connections and disconnections, by allowing migrations at one time, and preventing them from taking place in the reverse direction at a subsequent epoch, have been the chief factors which have resulted in the present very remarkable difference in the faunas of different parts of the globe. And it is solely due to such changes that many of the lower types of mammalian life, like the marsupials of Australia, and the lemurs and insectivores of Mada- gascar, have been preserved at the present day ; their insulation having afforded them protection from the invasion of the larger and more specialised mammals of other parts of the world, by which they would inevitably have been swept away, had the two groups ever come in contact. It is in regard to these migratory movements of animals and changes in the land-surface of the globe that zoology and geography are brought into such close relationship ; the former science sometimes helping to explain the alterations that have taken place in the contours of the land, while in other cases the present distribution of the land explains the past history of the animals by which it is inhabited. To understand rightly the present distribution of animals, it is, im ortance however, essential to study their past history as of Paiaeonto- recorded by the preservation of their fossilised remains in the strata of the earth's crust ; as with- out such history it would be quite impossible to grasp the reason of many apparent anomalies in their present distribution. How, for instance, without the aid of palaeontology would it be possible to understand how it came about that tapirs are now found only in tropical America and the Malayan countries, or that marsupials occur solely in America and Australasia at the present day ? And here it may be well to mention that the science of geographical distribution depends essentially upon a belief on the part of the student that all animals are genetically connected one with another, and that the existing forms have originated from earlier kinds by some mode of evolution. Were this belief not accepted, the whole science of distribution would fall to pieces ; as if animals were separately created, there would be nothing calling for special explanation in the fact of tapirs being restricted to the two areas mentioned. I.] IMPORTANCE OF PALEONTOLOGY. To those readers who may not be geologists, the following table of the leading divisions into which the strata of the earth's crust have been divided will probably be advantageous. Com- mencing with the highest beds, the series will run in descending order as follows, viz. : — Tertiary. PLISTOCENE. — Cavern and River Deposits. PLIOCENE. — The "Crags" of the East Coast. MIOCENE. — (Eningen beds of Baden. OLIGOCENE. — Gypsum of Paris Basin; Phos- phorites of Central France. EOCENE. — London Clay. Secondary. CRETACEOUS. — Chalk, Upper Greensand, Gault, Lower Greensand, and Wealden. JURASSIC. — Purbeck beds, Portland series, Kim- meridge Clay, Coral Rag, Oxford Clay, Great Oolite, Stonesfield Slate, Inferior Oolite, Lias. TRIASSIC. — New Red Sandstone of Cheshire. Paleozoic. PERMIAN. — Red Marls. CARBONIFEROUS. — Coal-Measures and Moun- tain-Limestone. SILURIAN. ORDOVICIAN. , « CAMBRIAN. Older rocks of Wales and the Lake-District. Before attempting to draw any conclusions as to the former configuration of the surface of the earth from the ,. . .,,...,... Inequality in distribution of the animals now inhabiting its dif- the Ages of dif- ferent countries, it is essential to understand that the different classes into which vertebrate animals are divided (and these only will be taken into consideration in the present volume) have a very different past history; the lower groups, such as fishes, reptiles, and amphibians being much older types than mammals and birds, and having attained their maximum development at a time when the two latter formed but a small minority of the earth's population. There is a considerable probability that at least a very large proportion of the animals that have populated the globe in 8 INTRODUCTORY. [CHAP, the later geological epochs originated high up in the northern hemisphere, if not, indeed, in the neighbourhood of the pole itself (which is known to have enjoyed a genial climate during the Tertiary period), and that they gradually migrated southwards in a series of waves, probably under pressure of the development of new and higher types in high latitudes; and it is to such southerly migrations that the present marked differentiation of the fauna of different parts of the earth's surface is chiefly due. Whether such a northerly origin held good for the terrestrial life of the Secondary epoch, there are no means of determining ; but it would appear that the higher animals (which were chiefly reptiles) of that epoch were very similar throughout the world, and that the differentiation of faunas had scarcely, if at all, commenced. Instances of this are afforded, as noticed in the sequel, by the occurrence of an identical genus of mammal (Tritylodori) in the lower Secondary rocks of Europe and South Africa ; as well as by the close alliance between the dinosaurian reptiles from the Jurassic rocks of Europe, North America, Argentina, India, and Madagascar (the genera being in some cases identical), and likewise between the anomodont reptiles of the Trias of Europe and the early Secondary rocks of South Africa and India. Reptiles belonging to orders still existing, such as crocodiles and chelonians (tortoises and turtles), had already attained a high degree of development in the Eocene division of the Tertiary period, when many genera now living had already made their appearance, whereas at that time the mammals were quite differ- ent from the modern forms. At the same time the side-necked tortoises (Pleurodira) were the dominant forms in the northern hemisphere, whereas they have now all migrated to southern lands, their place in the north being taken by the more specialised S-necked group (Cryptodira). This, however, is not all, for the rhynchocephalians, of which the sole existing representative is the New Zealand tuatera (Sphenodori), attained their maximum development in the northern hemisphere during the early part of the Secondary epoch, and their southern migration must have taken place during some portion of the same period. The palaeontological history of amphibians is still very imperfectly known, but since the group as a whole is an ancient one, the I.] DISTRIBUTION OF GROUPS DISSIMILAR. 9 migrations of the earlier forms must likewise probably have taken place at an early epoch. With mammals the case is very different. The earliest known forms, which date from the Triassic and Jurassic rocks, are chiefly marsupials and forms apparently allied to the monotremes, and it is probable that most of the descendants of these, as is more fully indicated in the sequel, migrated southwards during the early part of the Tertiary epoch, to find in Australasia a refuge from the competition of higher forms. Of the higher placental mammals, none of the modern types make their appearance before the Oligocene and Miocene periods, while many do not antedate the Pliocene. Their southern migrations accordingly took place later on in the Tertiary period, one of the earliest movements being the wandering of lemuroids, insectivores, and civet-like carnivores into South Africa and Madagascar. On the other hand, many other higher types, such as the hippopotami, giraffes, and antelopes, which were abundant in Europe and southern Asia during the Pliocene, only left their more northern homes to find a permanent abiding place in Africa at a very late epoch in the earth's history. Although the glacial epoch probably had a large share in the southern movements of the later Tertiary mammals, some cause with which we are unacquainted would appear to have been the impelling power at earlier epochs. But be this as it may, it is quite evident that a continuous series of waves of migrations of animal life has taken place throughout a very long portion of the earth's history. Similar migrations are also evident in the case of birds (which are likewise a modern group), many forms, such as secretary-birds and trogons, now exclusively southern in their distribution, being represented in Europe during the middle part of the Tertiary period. From this inequality in the ages, and consequently in the date of migration, of different groups of animals, it is manifest that there will be great differences in the present distribution of such groups; and hence it different . . .. , Geographical will be evident that zoological provinces indicated Distribution. by one group will not hold good for others. Notable instances of this are afforded by the very different divisions into which the globe is divided by those who take mammals, reptiles, 10 INTRODUCTORY. [CHAP. or amphibians as their standards. Between birds and mammals, as might have been expected from the comparatively recent high development of these groups, there is much greater accord. Birds, however, differ from mammals (with the exception of bats) in their power of flight, which enables many of them to cross wide ocean-tracts, and therefore renders them less valuable as indicative of the changes that have taken place in the distribution of land and water than the latter, which, as a rule, require direct means of land transit for their wanderings. Excluding man (and for the most part bats) and likewise the aquatic forms, such as seals, whales, and porpoises, o/ Mamma" niammals are the animals best adapted for parcelling in Geographi- out the globe into zoological provinces for two chief cal Distribu- 3 _. . , r , . , , tion. reasons. Firstly, they form a group which only at- tained its maximum development at a comparatively late epoch of the earth's history; and, secondly, their movements are mainly limited by the extent of the land-surfaces of the globe which were in actual communication at the time of such migra- tions. Consequently they afford the safest and truest indications of the latest changes which have taken place in the distribution of land and water. As reference in the following pages will constantly have to be .fi made to the various groups of mammals, it will be tion of Mam- well to give a list in this place of the chief ordinal and subordinal groups into which the class is divided ; such as are now extinct being indicated by an asterisk. The list stands as follows, viz. : — i. Order PRIMATES. — Apes, Monkeys, and Lemurs. 1 Suborder ANTHROPOIDEA. — Apes and Monkeys. 2 „ LEMUROIDEA.— Lemurs. ii. Order CHIROPTERA. — Bats. 1 Suborder MEGACHIROPTERA. — Fruit-bats. 2 ,, MICROCHIROPTERA. — Insectivorous Bats. iii. Order INSECTIVORA. — Insectivores. 1 Suborder DERMOPTERA. — Flying-Lemurs. 2 ,, INSECTIVORA VERA. — Shrews, Moles, Hedgehogs, etc. I.] CLASSIFICATION OF MAMMALS. II iv. Order CARNIVORA. — Carnivores. 1 Suborder CARNIVORA VERA. — Cats, Dogs, Bears, Weasels, etc. 2 „ PINNIPEDIA. — Seals and Walruses. 3 „ *CREODONTA. — Hyanodon, etc. v. Order RODENTIA. — Rodents. 1 Suborder SCIUROMORPHA. — Squirrels, Marmots, and Beavers. 2 „ MYOMORPHA. — Dormice, Mice, and Jer- boas. 3 „ HYSTRICOMORPHA. — Porcupines, Agutis, Cavies, etc. 4 ,, LAGOMORPHA. — Picas and Hares. vi. Order UNGULATA. — Hoofed Mammals. 1 Suborder ARTIODACTYLA. — Antelopes, Camels, Pigs, etc. 2 ,, PERISSODACTYLA. — Horses, Tapirs, and Rhinoceroses. 3 ,, *LITOPTERNA. — Macrauchenia, etc. 4 „ *ASTRAPOTHERIA. — Astrapotherium, etc. 5 „ *PYROTHERIA. — Pyrotherium. 6 „ *TOXODONTIA. — Toxodon, etc. 7 ,, HYRACOIDEA. — Hyraces. 8 ,, PROBOSCIDEA. — Elephants and Mastodons. 9 ,, *AMBLYPODA. — Coryphodon, etc. vii. Order SIRENIA. — Manatis and Dugongs. viii. „ CETACEA. — Whales and Porpoises, ix. „ EDENTATA. — Edentates. x. ,, EFFODIENTIA. — Aard-varks and Pangolins. xi. ,, MARSUPIALIA. — Pouched Mammals. 1 Suborder DIPROTODONTIA. — Kangaroos, Phalangers, and Wombats. 2 ,, POLYPROTODONTIA. — Dasyures, Bandicoots, etc. xii. Order MONOTREMATA. — Egg-laying Mammals, xiii. „ *MULTITUBERCULATA. — Pldgiciulax, Tritylodon, etc. 12 INTRODUCTORY. [CHAP. These orders are further subdivided into families and genera. In regard to the number of the latter of these, there is still a con- siderable difference of opinion among naturalists ; but in the present work those adopted in Flower and Lydekker's Study of Mammals1 will be in the main adhered to, with such corrections and additions as are necessary owing to recent emendations in nomenclature or to the discovery of new forms. Omitting from consideration the purely aquatic and volant members of the class, the most effectual barriers to Dispersal of the dispersal of mammals are formed by channels Mammals. their being crossed by swimming. And it is this inability to traverse any extent of water that renders what are known as oceanic islands practically devoid of all mammalian life, with the exception of a few bats and small rodents ; the latter animals having apparently some means of dispersal not common to other members of this class. Oceanic islands, it may be explained, are such as rise from great depths in the ocean, and are composed, almost invariably, either of volcanic rocks or of coral. They show, for the most part, no decisive evidence of having been connected with any continental land, and thus have never been enabled to receive a mammalian fauna2. In marked distinction to these are the so-called continental islands, such as Madagascar and Great Britain, which, both from the evidence of their mammalian fauna and their geological conformation, have indubitably been in direct communication with the adjacent continent at no very distant epoch. As a rule, the channels between such islands and the mainland are comparatively shallow, so that a moderate degree of upheaval would place the two in direct connection. The relative depth of the channel between two islands, or between an island and a continent is indeed of much more im- portance in regard to the dispersal of mammals than is its width. This is best exemplified by the well-known case of "Wallace's line " in the Malayan archipelago ; that name being applied to 1 London, 1891. 2 There is a possibility that some oceanic islands may have been connected with continents, and that their original mammalian fauna has been destroyed by submergence. I.] BARRIERS TO DISPERSAL. 13 the narrow strait separating the islands of Bali and Lombok, and its northward continuation, the Makassar Strait, dividing Borneo from Celebes. Although the two former islands are extremely close together, while Celebes is much less widely separated from Borneo than is the latter from Sumatra, yet the faunas of Lombok and Celebes are markedly distinct from those of the islands lying to the north and west of Wallace's line. Soundings show that the Makassar Strait, and likewise the Bali-Lombok Strait are of greater depth than the channels separating the other islands of the archi- pelago; and consequently that Wallace's line indicates a very old barrier which has long been impassable to the majority of mammals. That continental islands have received the great bulk of their mammalian fauna by means of a more or less complete land- connection with the mainland, is perfectly evident. Nevertheless, there are cases where certain mammals have crossed the inter- vening channel, either by swimming, or by having been carried across on natural rafts of some kind ; an instance of this nature being exemplified by the occurrence of an African type of pig in Madagascar. It thus becomes a question of considerable interest to ascertain what stretches of sea large mammals are capable of crossing. It is stated that the jaguar has been known to swim across the Rio de la Plata, which at its mouth is something like eighty miles across ; and a polar bear has been observed swim- ming at a distance of twenty miles from land in Bering Strait. The tiger frequently crosses the narrower channels in the Sandar- bans of Lower Bengal; and both deer, pigs, and elephants are good swimmers. The latter animals have, indeed, been known to swim for six hours at a stretch, and, with a rest, for upwards of nine ; but their rate of progress is extremely slow, and probably exceeds but little, if at all, a mile an hour. The Palk Strait, which is considerably less than forty miles in width at its narrowest part, has formed an effectual barrier to the passage of the tiger from India into Ceylon; and it may accordingly be assumed that about twenty miles is the utmost limit which mammals are likely to cross by swimming, even when favoured by currents. Such passages as these must, however, be of very rare occur- rence, for a terrestrial mammal is not likely to take it into its head 14 INTRODUCTORY. [CHAP. to swim straight out to sea in an unknown direction. Moreover, supposing a mammal new to a particular island to have arrived there by swimming, unless it happen to be a pregnant female, or unless another individual of the same species but of the opposite sex should arrive soon after (a most unlikely event), it would in due course die without being able to propagate its kind. And even if it should happen to be a pregnant female, there would be no certainty that its offspring, if but one in number, should be of the opposite sex to its parent. Accordingly, it would seem that the population of islands by mammals that have arrived by swim- ming must be a very rare event indeed. Rafts may be of more importance. Mr Aplin, in the Proc. Zool. Soc. for 1894, mentions that jaguars and pumas are frequently transported by them from one side of the Rio de la Plata to the other; and in the rainy season many are to be seen off the northern coast of Borneo, some of which may be as much as thirty yards in length. Still such rafts are not likely to cross straits, except when there is a current setting from one bank to the other. Before leaving this part of the subject, it must be mentioned that the degree of difference between the fauna of an island and that of the adjacent continent, or between the faunas of two islands, affords a most important clue as to the relative date of the land-connection between them. Madagascar, for instance, has a mammalian fauna which although clearly derived from Africa, is yet so different from that now inhabiting the mainland, that the land-connection between the two must have been broken up at an epoch comparatively remote. Ceylon and India present a condition in respect of their faunas intermediate between the last and that of Great Britain as compared to the Continent. In the latter instance, with the exception of a single Irish weasel, all the mammals are identical with species now inhabiting the Continent, thus proving that the connection has been a comparatively recent one ; although the peculiar weasel indicates that the separation between Ireland and Britain has been of sufficient duration to admit of the develop- ment of a distinct specific type in the former country. Although large rivers like the Amazon and La Plata un- doubtedly form serious barriers to the migration of mammals, yet these are not so insuperable as might at first sight be supposed, I.] BARRIERS TO DISPERSAL. 15 owing to the fact that in districts where vegetation is luxuriant, huge natural rafts are formed by the trunks of trees intermingled with vegetable matter, upon which numbers of animals may be borne down stream, and thus transferred from one bank to the other. Nevertheless, in treeless districts, or near their mouths, large rivers afford absolutely impassable barriers to the movements of mammals. In South America, for instance, even such an aquatic creature as the carpincho, or capibara (ffydrochcerus) has been unable to cross the Plata river from Uruguay into the Argentine, while, conversely the viscacha (Lagostomus) of Argentina is prevented by the same river from reaching Uruguay. Deserts are, perhaps, even more impracticable than rivers ; the Sahara — which was long supposed to have been the site of an ancient sea, although it has really been a desert since very remote ages — having apparently formed a barrier preventing the fusion of the mammals of North Africa with those to the south of that tract since at least the Pliocene epoch. It must not, however, be sup- posed that what are desert-tracts at the present day have always been such, the existence of a fossil chimpanzee in North-Western India during the Pliocene period indicating that the open sandy plains of the Punjab were at that time covered with dense tropical forests, and probably that the same was the case with parts of Syria and Arabia. Before taking leave of seas and deserts, it should be mentioned that in the polar regions ice may act in lieu of a land-connection to enable mammals to pass from one country to another. On this subject Dr Heilprin writes that " the reindeer is stated to cross the Bering Strait by way of the Aleutian Islands and the Frozen Sea, and in a somewhat similar manner the musk-ox finds its way to Melville Island; it is, however, somewhat singular that the last- named animal, despite its long ice-journeys, never manages to reach either the continent of Asia or Greenland." High mountain ranges form an effectual barrier to the migration of mammals, not only on account of the physical difficulties of crossing them, but likewise by the lowness of the temperature at great altitudes, coupled with the absence of proper food, being fatal to the existence of many. As already stated, however, mountain-ranges are much more far-reaching in their effects on such l6 INTRODUCTORY. [CHAP. migrations when, as in the Old World, their trend is from west to east, than when, as in America, they run from north to south, and thus do not interfere with the free movements of plain-dwelling animals on either side. In many instances the mammals inhabit- ing each of the isolated mountain-chains, as those of Europe, are to a great extent specifically identical one with another, not having had time to become modified into distinct species since they reached their present haunts at the close of the glacial period. Even among these, however, there are indications of the com- mencement of specific differences, the chamois of the Caucasus forming a variety differing somewhat from the typical Alpine race. Where the isolation has been longer, as in the case of the fauna of the highlands of Tibet, the difference is much more strongly marked ; the mammalian fauna in this instance being as peculiar and distinct as that of many ancient continental islands. Probably ever since man has existed in any numbers on the influence of gl°be he has been exerting a more or less strongly- ManonDistri- marked influence on the distribution of animals, either by destroying them, or by conveying them to countries or districts which are not their natural home. By the involuntary aid of man the common rat and mouse, which belong to a genus unknown in the New World, have been conveyed to every country in the globe ; while the rabbit has been carried to the Antipodes, where it has flourished and increased in an unprecedented manner. Cattle and horses have been introduced into South America, Australia, and other countries where they were naturally unknown, and by their rapid increase have shown that the absence of particular animals from particular districts is not necessarily due to their being unsuited to live there, but rather to the fact that they have been unable to find their way thither. The fallow-deer, again, has been imported from its Mediterranean home into England and other countries of northern Europe; while goats and pigs have been carried to a number of oceanic islands, where they have done irreparable harm in exterminating the native fauna and flora. In all these instances the fact of the introduction has always been more or less clearly known, and therefore no difficulty arises as to what are native and what are introduced forms. Very I.] MAN AND DISTRIBUTION. 1 7 different, however, is the case with the islands of the Malay Archi- pelago, where the natives, who have a wonderful facility for taming animals, have carried a species peculiar to one district or island to localities where it is quite unknown as a native ; and in consequence of this transportation and acclimatisation it is pro- bable that several mammals have been given a habitat to which they have not the most remote right. To the Malays is due the introduction of the small civet known as the rasse into Mada- gascar. Whether the dingo, or native dog of Australia, was intro- duced at an exceedingly remote era by the original colonisers of that island, or whether it is truly indigenous, is a question that will probably never be decisively answered. It is likewise quite impossible to say what part man may have played in the extermi- nation of the large mammals that inhabited Europe about the close of the glacial period, but it seems quite probable that he may have had a considerable share in their destruction. Be this as it may, the domestication of certain mammals has undoubtedly had the effect of destroying the wild race, as is remarkably ex- emplified by the two existing species of camel, of neither of which do we know the original habitat. The original European wild ox —unless, indeed, the half-wild cattle of the British parks be its direct descendants — has likewise disappeared at some unknown epoch owing to the hand of man. Although other mammals, such as the quagga (Equus quagga}, Burchell's rhinoceros (Rhinoceros simus), and the blaubok (Hippotragus leucophaus) have been almost or completely exterminated by human agency in South Africa, while the American bison has been practically swept away from its native prairies, yet in all these instances there is a more or less full record of the original range of the creatures. In other cases also mammals have been utterly exterminated by human agency from countries of which they were originally in- habitants, as is exemplified by the disappearance from the British Islands of the bear, the wolf, the beaver, and the wild boar within the historic period, although they still survive in other parts of their habitat. In these particular instances there is fortunately full evidence as to the former existence of these animals in Britain ; but it is highly probable that in more remote countries mammals have been exterminated without any record being left L. 2 1 8 INTRODUCTORY. [CHAP. of their existence, so that the full extent of their range, if they be surviving forms, can now never be ascertained. To quote the words of Dr Wallace, it is evident that in the present day we live in an impoverished epoch, so Extinction of the larger iar as the larger mammals are concerned, as com- pared with the Plistocene era ; this being true not only as regards the northern half of the Old World, but likewise North and South America, as well as Australia. From the northern half of the Old World have disappeared the mammoth, the elasmothere, the woolly and other species of rhinoceros, the sabre-toothed tigers, etc.; North America has lost the megalonyx and the Ohio mastodon ; from South America the glyptodonts, mylodons, the megalothere, and the macrauchenia have been swept away ; while Australia no longer possesses the dipro- todon and various gigantic species of kangaroos and wombats. In the northern hemisphere this impoverishment of the fauna has been very generally attributed to the effects of the glacial period, but although this may have been a partial cause, it can hardly be the only one. The mammoth, for instance, certainly lived during a considerable portion of the glacial epoch, and if it survived thus far, why should it have disappeared at the close? Moreover, all the European mastodons and the southern elephant (Elephas meridionalis] died out before the incoming of glacial conditions ; and the same is true of all the extinct elephants and mastodons of southern Asia. Further, a large number of English geologists believe the brick-earths of the Thames valley, which contain remains of rhinoceroses and elephants in abundance, to be of post-glacial age. As regards the southern hemisphere, it can hardly be contended that glacial conditions prevailed there at the same time as in the northern half of the world. It is thus evident that although a very great number of large mammals were exterminated (perhaps partly by the aid of human agency) at the close of the Plistocene period, when the group had attained its maximum development as regards the bodily size of its members, yet other large forms had been steadily dying out in previous epochs. And it would seem that there must be some general deep-seated cause affecting the life of a species with which we are at present' unacquainted. Indeed, as there I.] DISTRIBUTIONAL AREAS. 19 is a term to the life of an individual, what is more natural than that there should also be one to the existence of a species ? It still remains, indeed, to account for the fact that the larger Plistocene mammals had no successors in the greater part of the world, but perhaps this is in some way connected with the advent of man. Before coming to the consideration of the zoological divisions into which, from the present geographical distribu- tion of mammals, the world may be mapped out, it tionai Areas of is necessary to devote a brief space to the considera- Genera^ and tion of two other points ; the first relating to the relative size of the distributional area of genera and species, and the second to the permanency of ocean-basins and continents. As regards the first point, it appears to be true in the case of mammals (although not of all other groups) that every species has a continuous distributional area, except where this has been broken up by human destructiveness. It is not meant by this that every part of such area is inhabited by the particular species, as "station" renders this impracticable; but merely that the whole area is ranged over by the species in such spots as are suited to its particular mode of life. Great variation obtains, however, in regard to the size of such distributional area; and it will be obvious that the size of the area varies directly as the adaptability of the species to different climatic and other physical conditions. Perhaps the most important condition of all is the possibility of obtaining suitable food ; and in this respect carnivorous mammals are in a far better position than any other members of their class, since the kind of animal on which they prey is immaterial. This will readily account for the extensive geographical ranges enjoyed by the puma and the tiger, which, as stated on page 5, embrace almost every degree of latitude. Animals with such a wide dis- tribution are of but little use to the student of geographical distribution. Moreover, it will generally be found that species with a wide range belong to large genera having a still more extensive distributional area; this being markedly the case with the puma and the tiger ; the genus Felts being one of the largest in the class, and ranging over the whole world with the exception of Australasia. Such cosmopolitan genera are likewise almost valueless to the distributionist. 2 — 2 20 INTRODUCTORY. [CHAP. On the other hand, species with a small distributional area usually belong to small genera, of which they may be the only representatives. Instances of this nature are afforded by the panda (sElurus] and binturong (Arctictis) of the eastern Himalaya, and the parti-coloured bear (^Eluropus] and chiru antelope (Pan- tholops] of the Tibetan plateau. Although such single representa- tives of genera are highly important to the study of distributional zoology, of vastly greater importance are small genera having two or more species living in widely separated areas. Examples of such are to be found among the porcupines of the genus Atherura, of which one species is Malayan and the other two West and Central African ; in the mice of the genus Golunda, with one African and one Indian representative ; and likewise by the tapirs (Tapirus\ of which there is one Malayan, and several tropical American species. These examples of " discontinuous distribu- tion " among genera are of the very highest import to the science ; since they clearly indicate that some of the lands lying between its present disconnected distributional areas must have formerly been the habitat of the genus, and thus enable important conclusions to be drawn as to the former land-connections between such areas. Both in the case of the tapirs and of the brush-tailed porcupines, remains of extinct species have been discovered in the intermediate areas. Equally important are families, either large or small, which contain two or more closely allied small genera respectively con- fined to distant areas. As an instance of a large family containing such allied genera, may be cited the Viverridce, among which the true linsangs (Linsangd) are represented by several species from the Eastern Himalaya and the Malayan countries, while the closely-allied Poianais confined to West Africa. The chevrotains (Tragulidee), on the other hand, form a small family with a dis- continuous distribution ; one genus (Tragulus) being now Oriental, while the other (Dorcatheriuni) is West African. Here it is quite evident, of course, that the distributional area of the family must once have been continuous ; and, as a matter of fact, remains of both genera occur in the Pliocene of India, those of the latter being also found in the European Miocene. In other families with a discontinuous distribution, as in the I.] CENTRES OF EVOLUTION. 21 rodent family Octodontida, which is now mainly confined to Africa south of the Sahara, and Central and South America, the genera may be less closely allied, although sufficiently so to indicate a continuous distributional area, or rather a common centre of dispersal, at no very remote epoch. Allied families, with a small number of genera, severally con- fined to distant localities are likewise of the highest value in building up the former history of the globe. As examples of this nature may be cited the tree-shrews (Tupaiida) of the Oriental countries and the jumping shrews (Macroscelidida) of Africa on the one hand, and the Solenodontidce of the West India islands, and the Centetidce of Madagascar on the other. Such families must clearly have had a common centre of origin and dispersal ; the available evidence suggesting that in the case of the two former such centre was Europe. Although of far less common occurrence than among families or genera, discontinuous distribution in an order is perhaps of even more importance than either of the other cases, as it implies a greater interval of time since the original dispersal took place, and, therefore, carries back such conclusions as can be drawn in regard to former land-connections to a still earlier epoch. Among mammals the only instance of this nature is to be found in the marsupials1, of which two families are American (and mainly South and Central American), while all the others are confined to Australasia and some of the adjacent Malayan islands. In this case also there is abundant evidence of the wide distribution of the whole group in former epochs of the earth's history. This last instance leads on to the consideration of what may be termed "centres of evolution." In a previous paragraph it has been stated that, according to the " °f available evidence, a very large proportion, if not the whole, of the terrestrial mammalian life of the globe has originated in the northern hemisphere, from which it has spread southwards in a continuous successive series of waves. When, however, certain groups of mammals had once reached the more 1 In this work the Effodientia are separated from the Edentata; but when these are united, there is a second instance. 22 INTRODUCTORY. [CHAP. remote parts of the southern hemisphere, where they were free from the competition of the higher forms, and met with favourable conditions, they seemed to take a new lease of life, and attained a fulness and variety of development which they had never reached before. As a rule, more or less complete isolation has been a dominant feature of this development; of which the best and most striking instance is that of the marsupials in Australasia. That area may accordingly be called the marsupial evolutionary centre. Scarcely less striking is the instance of the edentates (of which the original derivation is unknown) in South America, where, in company with certain peculiar extinct groups of ungulates, they attained an extraordinary development, both as regards the number of specific, generic, and family types, and likewise in respect of the bodily size of some of its members. This second area may be termed the evolutionary centre of the edentates. A third great centre is constituted by Europe, Asia, and North America, which appear to have been the main developmental centre of the higher mammals, and may accordingly be named the placental evolutionary centre. Two other minor centres are respectively indicated by Madagascar and Africa south of the Sahara : the former as being the headquarters of the lemurs, may appropriately be spoken of as the lemuroid centre, while the great development of the antelopes in Africa suggests the name of the antelopine evolutionary centre for that continent. The circumstance that throughout the greater part of North America and Europe a very large proportion of the of Conine ntsy continents are built up of sedimentary strata of and Ocean- marine origin, naturally led geologists in the early Ba.si.ns. . . . 1 1*1 days of their science to the conclusion that every part of the land had at one time been deep ocean, and every stretch of ocean dry land. More careful study led, however, to the belief that this idea was not founded on fact, and that although it was perfectly true that what are now continents had been many times under the sea, yet that such areas had never formed abyssal ocean-depths ; and, conversely, that such ocean- depths had never been dry land. In addition to many other lines of evidence, this view of the permanency of continents and ocean- basins is strongly supported by the circumstance that nearly all I.] CONTINENTS AND OCEAN-BASINS. 23 oceanic islands are either of volcanic or coral origin, and do not contain sedimentary rocks; and also that deposits analogous to those laid down in the deepest ocean beds are generally wanting from among the sedimentary series of rocks of which the continents and islands are composed. A further argument was afforded by the discovery that the greater portion of peninsular India and South Africa has been dry land since the Palaeozoic epoch. As is so commonly the case in similar instances, the promulga- tors of the doctrine of the permanency of continents and ocean- basins pressed their hypothesis too far; and it is now evident that although the doctrine is true as a whole, and more especially as regards the later stages of the earth's history, yet it requires very considerable modification from the original form in which it was advanced. In the first place, it has been shown that crystal- line granitic and gneissic rocks occur in the Seychelles, which were formerly regarded as true oceanic islands ; and, secondly, deep-sea deposits have been discovered in the West Indies and the Solomon Islands. Moreover, various lines of evidence indicate that during the Jurassic and Cretaceous epochs there was a continuous land- connection between Africa (by way of Madagascar and the Seychelles) and India; while at some time in the Secondary era, in the opinion of Drs Neumayr and Blanford, South America and South Africa were in communication across the South Atlantic. The latter connection appears, indeed, to have been a survival from an older Palaeozoic girdle of land which, from the evidence of fossil floras, seems to have existed in low latitudes round nearly three-quarters the circumference of the globe, and which was cut oft" from the land to the north. There is, moreover, the possibility of a Tertiary connection of Australia with Patagonia by way of Polynesia, to which allusion is made in the third chapter. Then, again, the recent investigations of Dr J. W. Gregory1 on the fossil corals of the West Indies have afforded strong support to the view that the Atlantic is of comparatively recent origin. After referring to the remarkable resemblance between the existing fauna of the West Indian seas and that of the Miocene deposits of the Mediter- ranean basin, Dr Gregory2 observes that the sea-urchins, or 1 Quart. Joitrn. Geol. Soc. vol. LI. pp. 255 — 312 (1895). 2 Ibid., pp. 306, 307. 24 INTRODUCTORY. [CHAP. echinoderms, yield still more conclusive evidence. "As I have previously pointed out," he writes, "the intimate affinity between those of the West Indies and the Mediterranean can only be explained by the assumption of the existence of a shallow-water connection across the Central Atlantic in — at latest — Miocene times. That the fauna did not follow along the shores of the North Atlantic basin, is shown by its absence from the northern Miocenes of Europe and North America. The evidence now adduced from the fossil corals of Barbados lends support to this view, as showing that the West Indian fauna is only a fragment of that of the Mediterranean Miocene, and has received nothing from the Pacific. This is in full agreement with Prof. Suess's theory that the Atlantic is of comparatively recent geological age, and arose by the gradual enlargement of two bays which ran north and south from a sea that once extended across the Mid-Atlantic from Europe to America, including both the Mediterranean and the Caribbean Sea." The question of the southward extension of America, Africa, and Australia to join the Antarctic continent during Tertiary times is alluded to in the sequel. Summing up the evidence in regard to the permanency of oceans and continents, Dr Blanford1 several years ago observed " that whilst the general permanence of ocean-basins and conti- nental areas cannot be said to be established on anything like firm proof, the general evidence in favour of this view is very strong. But there is no evidence whatever in favour of the extreme view accepted by some physicists and geologists that every ocean-bed now more than 1000 fathoms deep has always been ocean, and that no part of the continental area has ever been beneath the deep sea. Not only is there clear proof that some land-areas lying within continental limits have at a comparatively recent date been submerged over 1000 fathoms, whilst sea-bottoms now over 1000 fathoms deep must have been land in part of the Tertiary era, but there are a mass of facts both geological and biological in favour of land-connection having formerly existed in certain cases across what are now broad and deep oceans." 1 Appendix, No. 8, p. 107. I.] ZOOLOGICAL REALMS. 25 Although much previous work had been done on the subject, the first real attempt to divide the land-areas of the , , . i-i • • Zoological globe into zoological provinces, or regions, was made Realms and by Dr P. L. Sclater1 in 1858. According to this Regions' scheme, which was mainly based on the study of Passerine birds, the world was parcelled out into the following six zoological regions, viz.: — 1. Palczarctic ; Europe, Northern Africa, Northern and Cen- tral Asia. 2. Ethiopian ; Africa south of the Atlas, and Madagascar. 3. Indian, renamed Oriental by Dr Wallace; India, South- eastern Asia, and part of the Malay Archipelago. 4. Australian ; Australia, with New Guinea and the adjacent islands, New Zealand, and Polynesia. 5. Nearctic ; America as far south as Mexico. 6. Neotropical ; Central and South America, with the West Indies. This scheme, which has been adopted and developed in the brilliant writings of Dr Wallace, has the important merit that it coincides to a great extent with the leading geographical divisions of the globe. It has, however, the serious drawback that it gives no greater rank to Australasia and South America than to the other divisions ; whilst the remarkable difference between the fauna of Africa and Madagascar is overlooked. Further, the northern parts of America are widely separated from those of Europe and Asia to which they are faunistically extremely close. It should be added that in Dr Sclater's scheme the first four regions, or those belonging to the Old World, were brigaded together under the title of PAL^OG^EA, while the two last, or New World regions, were bracketed as NEOG^EA. The next important classification was one propounded in 1868 by Professor Huxley2, who, basing his conclusions on the distri- bution of the game-birds, divided the world into a northern and a southern division, taking the name of ARCTOG^EA for the former, and NOTOG^EA for the latter ; Notogaea being further sub-divided into a Novo-Zelanian (New Zealand), Australian, and Austro- 1 Appendix, No. 26. '2 Ibid., No. 18. 26 INTRODUCTORY. [CHAP. Columbian region, the latter being equivalent to the Neotropical of Sclater. Six years later, Dr Sclater1, who had by this time turned his attention to the distribution of mammals, proposed to group the regions he had previously named under three larger divisions, making a fourth division for New Zealand and Polynesia. This scheme is as follows, viz. : — I. ARCTOG^EA. — Palsearctic, Nearctic, Oriental, and Ethiopian regions. II. DENDROG^A. — Neotropical region. III. ANTARCTOG^EA. — Australian region (exclusive of New Zealand and Polynesia). IV. ORNITHOG^EA. — New Zealand and Polynesian region. So far as mammals are concerned, this scheme was a great advance on the first one, although the distinctness of Madagascar was not recognised, while the Palasarctic and Nearctic regions were still maintained. Most of the names for the major divisions are, however, open to objection. In 1878 Dr Heilprin2, who does not employ these larger groups, proposed, after a suggestion of Professor A. Newton, to unite Dr Sclater's Palsearctic and Nearctic regions under the common title of the Holarctic region ; separating, however, from the former a "transitional" Mediterranean region, and from the latter a similar Sonoran region. A further step was made in 1890 by Dr Blanford8, who pro- posed the following scheme, viz. : — I. Australian region. II. South American region. III. Arctogaan region ; this being divided into Malagasy, Ethiopian, Oriental, Aquilonian ( — Palaearctic and northern part of Nearctic), and Medio-Columbian ( = Sonoran). Several other minor modifications have been suggested from mammalian evidence, Dr Allen4 in 1892 reviving the view that the Oriental and Ethiopian regions should be united, under the name of the Indo-African ; but the next most important memoir is that 1 Appendix, No. 27. 2 Ibid., No. 17. 3 Ibid., No. 8, p. 76. 4 Ibid., No. 2. I.] REALMS AND REGIONS. 2J of Dr Hart Merriam1 in 1892, whose views are fully discussed in the sequel. It will accordingly suffice for our present purpose to say that in 1893 an anonymous writer2 proposed to take the terms NOTOG^EA, NEOG^A, and ARCTOG^A to indicate the three major divisions of Dr Blanford's classification ; the same terms being used by Mr W. L. Sclater3 in a nearly similar sense. The following scheme is the one adopted in the present volume, viz. : — I. NOTOG^EIC REALM. — i. Australian Region. 2. Polynesian Region. 3. Hawaiian Region. 4. Austro-Malayan Region. II. NEOG^EIC REALM. — Neotropical Region. III. ARCTOG^EIC REALM. — i. Malagasy Region. 2. Ethiopian Region. 3. Oriental Region. 4. Holarctic Region. 5. Sonoran Region. It will be noticed that the three realms correspond to the three great evolutionary centres of mammals alluded to in an earlier page. It may be added that, in a work expressly devoted to the geographical distribution of mammals, it will be unnecessary to allude to the schemes proposed on the evidence of other groups of animals, and we may accordingly proceed forthwith to the con- sideration of the distinctive features of the various realms and regions here adopted. 1 Appendix, No. 19. 2 Ibid., No. 4, p. 289. 3 Ibid., No. 28. CHAPTER II. THE NOTOG^IC REALM. Definition and Characters of the Realm — Australian Region — Monotremes— Marsupials — Rodents — Carnivores — Ungulates — Bats — List of Australian and Papuan Genera — Polynesian Region — Hawaiian Region — Austro- Malayan Region — Palaeontological History of Marsupials — How Australia received its Fauna. THE term Notogaea was first proposed, as stated in the preced- ing chapter, by Professor Huxley1, to include not only the Australian region of Dr Sclater, but likewise the Neotropical region (Austro-Columbia) ; but an anonymous writer2 appears to have been the first to restrict it to the former of these areas3. This view, as being, on the whole, the most convenient, is adopted here; and the Notogasic realm may accordingly be taken as the first of the three primary zoological divisions of the globe, and as equiva- lent to the Australian region of Drs Sclater and Wallace. Accord- ing to the latter writer4, " its central and most important masses consist of Australia and New Guinea, in which the main features of the region are fully developed. To the north-west it extends to Celebes, in which a large proportion of the Australian characters have disappeared, while Oriental types are mingled with them to such an extent that it is rather difficult to determine where to locate it. To the south-east it includes New Zealand, which is in some respects so peculiar that it has even been proposed to con- stitute it a distinct region. On the east it embraces the whole of Oceania [Polynesia] to the Marquesas and Sandwich Islands, 1 Appendix, No. 18. '2 Ibid., No. 4. 3 The term Antarctogsea has been proposed by Dr Sclater (Appendix, No. 27, p. 214), for this area, but it is not a happy one. 4 Appendix, No. 32, vol. i., p. 387. CHAP. II.] THE NOTOG/EIC REALM : ITS CHARACTER. 29 where a very scanty and often peculiar fauna must be affiliated to the general Australian type." To the north-east the line of de- marcation of the realm from the Oriental region of Arctogsea has been finally fixed at the deep channel separating the islands of Celebes and Lombok on the one side from those of Borneo and Bali (at the extremity of Java) on the other ; this division being now well known under the name of " Wallace's line." All writers are, however, by no means agreed as to the right of the whole of the area thus indicated to form a single zoological division. Before the publication of Dr Wallace's great work, Professor Huxley had proposed to separate New Zealand as a region of equal rank with his Australasian region. At a later date Professor Heilprin1 suggested that the Australian realm should include only Australia, Tasmania, New Guinea, with the smaller Papuan islands, and New Zealand ; Polynesia, including all the islands lying to the east of the Coral Sea, being raised to the rank of a distinct realm (the Polynesian), while the Austro-Malayan islands were regarded as forming a transitional tract between the Australian realm and what is here termed the Oriental region. In this connection it may be well to notice that the Austro-Malayan sub-region of Dr Wallace is by no means coterminous with the Austro-Malayan transition-tract of Heilprin, the former including, and the latter excluding New Guinea. So far as mammals alone are concerned, Notogaea is widely separated from the whole of the rest of the world by being the sole habitat (both now and in the past) of the typical diprotodont marsupials and the monotremes; although it must not be sup- posed that either of these groups is distributed over the entire area. As a matter of fact, apart from introduced rodents, Polynesia is devoid of mammalian life with the exception of bats and a rat 2, while New Zealand has but two representatives of the former group, and a rat which may or may not be indigenous. But wherever we meet with a fully developed mammalian fauna, as in the transitional Austro-Malayan islands, there a certain number of marsupials are met with, although the monotremes 1 Appendix, No. 17. 2 Mus exulans, see Proc. Zool. Soc. 1895, p. 338. 30 THE NOTOGyEIC REALM. [CHAP. are restricted to Australia, with Tasmania ; and New Guinea, with the adjacent islands, such as the Aru group. The Notogaeic realm, as denned above, may be conveniently divided into four distinct regions, as follows. Firstly the Australian region, comprising Australia, Tasmania, New Guinea and the adja- cent Papuan islands; characterised by marsupials and monotremes forming by far the predominant element in the mammalian fauna. Secondly the Austro-Malayan region, embracing Lombok, Celebes and the other islands lying between them and the Australian region; this area being characterised by the absence of monotremes and by the marsupials (all of which belong to the diprotodont division of the order) forming only a small minority of the mammalian fauna. Thirdly, there is the Hawaiian region, including only the Sandwich Islands; and, lastly, the Polynesian region, which maybe taken to include all the islands, save those last named, lying to the eastward of the Coral Sea, together with New Zealand, and is characterised by the general absence of terrestrial mammals. There is some difficulty in deciding whether the islands of the Solomon group should be included in this region, or classed with the Papuan division of the Australian region, seeing that, in addition to a considerable number of bats, they have four species of mice, and one diprotodont marsupial (Phalanger)1. When, however, the poverty of this fauna as compared with that of Papua is taken into consideration, and it is also borne in mind that Mr C. Hedley 2 has come to the conclusion that the Solomon Islands, together with the New Hebrides, New Caledonia, Fiji, and Norfolk Island, are closely connected by means of their flora with New Zealand, and have but little in common with Australia and New Guinea, it seems preferable to include the former group in the Polynesian region. On the same grounds, New Zealand is regarded, in accordance with the views of Heilprin, as also forming a portion of the same zoological region, and not as the representative of a separate region by itself. The fauna of the Solomon Islands has doubtless been derived directly from that of the Duke of York group, which clearly belongs to the same region as New Guinea, and shows a much more strongly marked Papuan facies, having three species of mice, and four marsupials. 1 See Thomas, Appendix, No. 30. 2 Appendix, No. 16. II.] AUSTRALIAN REGION. 31 Although the northern half of Australia lies within the tropics, yet few portions of this great island present that luxuriance of vegetation which we are accustomed to associate with tropical scenery ; and large tracts of the interior, owing doubtless to the absence of elevated moun- tain ranges in the central districts, form arid sandy deserts more or less unsuited to the maintenance of animal life. The coast regions and the borders of the larger rivers are accordingly those where vegetation flourishes best ; the finest tracts of pasture-country, well supplied with water, lying to the east and south-east, and Victoria possessing a mountain range whose summits are perpetually clothed with snow. Mountains also occur in the dry and hot western districts. Although Tasmania enjoys moister conditions, Australia as a whole is characterised by the lack of water and the general dryness of its climate ; and it is probable that to this aridity the number of jumping animals, such as kangaroos, rat- kangaroos, and jerboa-rats, — now characteristic of this part of the region — is due, since such creatures are admirably adapted for traversing long distances in search of food and water. On the other hand, New Guinea, together with the Papuan islands, has a moist tropical climate, essentially different from that of Australia, but similar to the conditions obtaining in a large portion of the Austro- Malayan islands. Hence it is not to be wondered at that the mammals of New Guinea differ very markedly from those of Australia ; this being especially noticeable in the paucity of typical jumping kangaroos, and the proportionately large number of arboreal members of this group. Nevertheless the mammalian fauna of Queensland and North Australia exhibits a marked approximation to that of New Guinea, one species of kangaroo, as well as a cuscus, a striped phalanger (DaJylopsila), a flying phalanger (Petaurus), a pouched-mouse (Phascologale), and an echidna, being common to the two areas, and it is in these countries alone that tree-kangaroos are met with. From these resemblances in their faunas — and especially from the restriction of the monotremes to these two areas, — there can be no question as to the propriety of including Australia and New Guinea in the same zoological region, and thus separating the latter country from the Austro-Malayan region. THE NOTOG^IC REALM. [CHAP. The egg-laying mammals, or monotremes, constitute not only a distinct order (Monotremata), but likewise a Monotremes. . N separate sub-class (Prototheria); and are broadly distinguished from all other members of their class by laying eggs, from which the young are in due course hatched; as they are likewise by the milk-glands of the female opening on the surface of the skin by means of a number of minute perforations, without being furnished with nipples. The group is represented by three genera, FIG. i. THE DUCKBILL. (Ornithorhynchus anatinus.} one of which is widely different from the other two and forms a family by itself, while the latter constitute a second family. The duckbill {Ornithorhynchus anatinus}, as the single representative of the first family (OrnithorhynchidcB) is commonly termed, is an aquatic, somewhat mole-like, burrowing animal, easily recognised by the expansion of the muzzle into a broad duck-like beak covered during life with a sensitive skin, and also by the broadly webbed feet, of which the soles are naked and devoid of pads. II.] MONOTREMES. 33 Although in the adult the mouth is furnished only with horny plates, in young individuals the sides of the jaws are provided with three pairs of molar teeth, quite unlike those of any other living mammals. At the present day the duckbill is confined to Queensland south of latitude 18°, New South Wales, Victoria, South Australia, and Tasmania; it is represented by an extinct species from the Plistocene of Queensland, but otherwise the pakeontological record of the group is a complete blank. Just the same is the case with the echidnas, or spiny anteaters (Echidnidce], of which the only fossil remains known have been obtained from the superficial deposits of New South Wales. Terrestrial and fossorial in their habits, the echidnas differ from the duckbill in having the muzzle in the form of an exceedingly slender cylindrical toothless beak, furnished with an extensile worm-like tongue ; while the fur is thickly mingled with short spines, the tail being rudimental, and the unwebbed toes provided with extremely powerful claws. Of the two species, the common five-clawed echidna (Echidna aculeata] extends from south-eastern New Guinea throughout Australia to Tasmania; whereas the three-clawed echidna (Proechtdna1 bruijni} is restricted to New Guinea. With the exception of the Plistocene forms already alluded to, no fossil monotremes whatever are known to science. It is, how- ever, not improbable that certain extinct mammals from the Secondary and lower Tertiary rocks of Arctogaea, commonly termed Multituberculata, which will be more fully alluded to in the sequel, may indicate a second order of the sub-class Prototheria. Both the extinct and the living groups are, however, of a highly special- ised type, so that the one cannot apparently be regarded as ancestral to the other ; but if the presumed distant relationship between the two be substantiated, it will indicate that we are to look to a northern origin for the existing monotremes. The marsupials, which likewise represent both a separate order (Marsupialia) and a sub-class (Metatheria) Marsupials. oy themselves, differ from the monotremes by producing living young, and by the milk-glands of the female 1 It has recently been proposed to substitute the term Zaglossus, which is stated to be earlier, for this genus. L. 3 34 THE NOTOG^IC REALM. [CHAP. discharging their secretion by means of nipples. From the higher mammals (Eutheria) they are distinguished by the imperfectly developed condition of the newly-born young, and the absence of any prenatal connection between the vascular system of the foetus and the maternal parent by means of the organ known as the placenta. Very generally the young are carried about for some time after birth in a pouch situated on the abdomen of the parent, where they at first remain immovably fixed to the nipples, the milk being injected into their throats by the action of a special muscle. In the carnivorous and insectivorous forms the number of incisor teeth in the upper jaw usually exceeds the three pairs which form the general maximum limit in the higher mammals. A further peculiarity of the order is to be found in the replacement of the teeth. Instead of the whole or nearly the whole of the first, or milk-set of teeth in advance of the true molars or hinder cheek- teeth being replaced by a second set of permanent teeth, only one tooth is thus (and that by no means invariably) replaced. The tooth thus replaced was long regarded as corresponding to the last or fourth milk-molar of the higher mammals, while the apparently replacing tooth was identified with the last or fourth premolar of the same. From recent researches, however, it would seem that in reality this is not a case of true replacement, and that the tooth which makes its appearance late in life is a retarded premolar, representing the fourth in that series, while the replacing tooth is the fifth. Marsupials may be divided into two main sections or sub- orders, readily distinguished from one another by their dentition, both of which are represented in Notogaea. In the first of these, or Diprotodont sub-order, which is the more specialised of the two, the incisor teeth are separated by a gap from those of the cheek- series, and do not usually exceed three in number on each side of the upper jaw1, and in the lower jaw are generally reduced to a single pair, while the tusks, or canines (<:), are either small or wanting. In their habits the members of this section are more or less exclusively herbivorous. On the other hand, in the Polyprotodont mar- supials, all of which are mainly carnivorous or insectivorous in 1 The only exception to this occurs in the South-American forms. II.] MARSUPIALS. 35 their diet, the incisor teeth are numerous and pointed, the canines are large and well developed, and the whole of the anterior teeth form a series more or less nearly continuous with those on the sides of the jaws. The typical diprotodonts, or those in which two of the toes of the hind foot are enclosed in a common integu- ment1, are exclusively confined to the Notogaeic realm, where they attain their maximum development in the Australian region ; but the polyprotodonts and an aberrant group of diprotodonts are still represented in the Neogaeic realm, while during the Secondary and earlier part of the Tertiary period the former were widely FIG. 2. SKULL OF RAT-KANGAROO. (To exhibit Diprotodont type of dentition.} spread over Arctogaea. In the Australian region marsupials play the part of the eutherians of other regions, and show a remark- able diversity of external form and structure, adapting them to all modes of life with the exception of the aquatic. And it is fairly evident that within the limits of this region the diprotodonts were originally evolved from the more generalised polyprotodonts. Of the three existing family groups into which the Notogaeic di- protodonts are divided the first is the Macropodidce, or the kangaroos and their allies ; this being in some respects the most specialised group of all, and characterised by certain peculiar features in the skull and dentition. Among these the typical genus Macropus, in- cluding the true kangaroos, comprises a total of twenty-three species, out of which twenty are confined to Australia and 1 The term syndactylous is applied to this type of foot. 3—2 36 THE NOTOG^IC REALM. [CHAP. Tasmania, one (M. agilis) is common to Australia and Queens- land, and two others (M. bruijni and M. browni} are confined to New Guinea or the adjacent islands. Of the six species of rock- kangaroos (Petrogale), none are found out of continental Australia, and the same is true with regard to the three representatives of the nail-tailed wallabies (Onychogale), and likewise with the three hare-wallabies (Lagorchestes}. On the other hand, the three kinds of dorca kangaroo (Dorcopsis) are exclusively Papuan ; while of the climbing tree-kangaroos (Dendrolagus), three are from Papua and two from Queensland. The single species of banded wallaby (Lagostrophus) is Australian ; as are also the whole of the rat- kangaroos, forming the genera Potorous, Caloprymnus, Bettongia, and sEpyprymnus, and likewise the peculiar musk-kangaroo (Hypsiprymnodon), which serves to connect the other members of the family with the phalangers. Several of the existing representatives of the above-mentioned genera are found in a fossil state in the cavern-deposits of New South Wales and the Plistocene formations of Queensland, in addition to which there are likewise several extinct representatives of the genus Macropus, some of which considerably exceed the largest living forms in point of size. The same formations have also yielded the remains of three extinct genera, namely Palor- chestes, Procoptodon, and Sthenurus, all of which appear to have been allied to the wallabies, although some of the species were vastly larger than any existing kangaroo. Another, but very im- perfectly known genus Triclis, seems to have connected the musk-kangaroo so closely with the phalangers, that it is scarcely possible to draw any distinction between these two families. In the family of the phalangers (Phalanger idcz], which differs from the more typical representatives of the preceding by the more generalised characters of the skull, teeth, and limbs, there is an exclusively Australian form in the koala, forming the sole representative of the genus of the same name. Of the five species of cuscuses (Phalanger}, one is, however, common to northern Australia, New Guinea, and the Austro- Malayan Islands, while the other four are restricted to the two latter areas. The two species of true phalanger (Trickosurus) are, on the other hand, exclusively Australian ; while the ring-tailed II.] MARSUPIALS. 37 phalangers, constituting the genus Pseudochirus, are common to Australia and New Guinea. Another exclusively Australian type is to be found in the taguan flying phalanger (Petauroides) ; but of the non-volant striped phalangers (Dactylopsila) one species is common to Queensland, the Aru Islands, and New Guinea, while the second is exclusively Papuan. The true flying phalangers of the genus Petaurus include two Australian species, and a third, common to northern and eastern Australia and New Guinea and the adjacent islands. Leadbeater's phalanger (Gymnobelideus), which appears to be closely related to the ancestral stock from which were evolved the members of the last genus, is restricted to Victoria ; but the dormouse-phalangers of the genus FIG. 3. SKULL OF EXTINCT PHALANGER (Thylacoleo carnifex). Dromicia have both Australian and Papuan representatives, while the pen-tailed phalanger (Distachurus) is exclusively from New Guinea, and of the two pigmy flying phalangers (Acrobates), one is Australian and the other Papuan. Lastly, the aberrant long- snouted phalanger ( Tarsipes), representing a sub-family by itself, is confined to Western Australia. Remains of species belonging to some of the existing genera have been disinterred from the caves of New South Wales and the Plistocene deposits of Queens- land; while several more or less imperfectly known extinct generic types have been described. Among the latter, by far the most remarkable is Thylacoleo, which was a gigantic phalanger comparable in size to a large leopard, and distinguished by the great development of the last premolar tooth in each jaw. The 38 THE NOTOG^IC REALM. [CHAP. tooth in question has an elongated cutting-blade, adapted to work against its fellow in the opposite jaw with a scissor-like action, somewhat after the fashion of the carnassial teeth of a tiger ; but the other cheek-teeth were all relatively small, although the tusks were large. The giant among the marsupials was the extinct Diprotodon of the Australian Plistocene, a creature rivalling in size the extinct South American Megalotherium, and allied on the one hand to the kangaroos, and on the other to the phalangers. It was not, however, endowed with the leaping powers of the former, and doubtless walked on the ground in the ordinary manner, its toes having apparently been covered with structures intermediate between hoofs and nails. Nearly related, but likewise repre- senting a family by itself, is the somewhat smaller, but still gigantic Nototherium, which in the conformation of its limb-bones appears to approximate to the wombats, and may consequently have been, like those animals, of fossorial habits. The last Notogaeic family of the Diprotodont section is that of the wombats (Phascolomyidcz\ distinguished from all the preceding forms by the presence of only a single pair of incisor teeth in both the upper and lower jaws ; canines being absent, and the whole dentition thus curiously simulating that of the rodents among the higher mammals. All the three existing species, which are included in the single genus Phascolomys, are confined to Australia and Tasmania; and, except certain extinct species belonging to the same genus, the only other member of the family is the extinct Phascolonus (Sceparnodoti) from the Australian Plistocene, distinguished by the peculiarly flattened and chisel- like form of the upper incisor teeth. This, the only known species, attained much larger dimensions than either of the existing wombats. The Polyprotodonts likewise include three existing families found within the limits of the Australian region, none of the mem- bers of which stray either into the Austro- Malay an or Polynesian regions, although the separate family of the opossums (Didelphyidcz} inhabits the New World. In the family of the bandicoots (Pera- melidtz}, the two species of rabbit-bandicoot (Peragale) are ex- clusively Australian, whereas the true bandicoots (Perameles] have both Papuan and Australian representatives ; the third genus II.] MARSUPIALS. 39 ( Chceropus\ which includes only the pig-footed bandicoot, being confined to Australia. In the second family, or Dasyurida, the genus Thyladnus is now confined to Tasmania, but it was represented during the Plistocene period on the Australian mainland, where one species is stated to have been obtained from beds of Pliocene age. A similar distribution also obtains in the case of the genus Sarcophifas, now represented only by the well-known Tasmanian devil. FIG. 4. BANDED ANTEATER. (Myrmecobius fasciatus.) Although mainly Australian, the smaller animals known as dasyures (Dasyurus] have, however, a single Papuan representative ; while the pouched mice (Phascologale) are likewise common to the two areas, one of the species ranging from New Guinea to eastern and southern Australia. On the other hand, both the narrow-footed pouched mice (Sminthopsis) and the jumping pouched mouse (Antechinomys] are exclusively Australian. The same is the case 40 THE NOTOG^IC REALM. [CHAP. with the aberrant banded anteater (Myrmecobius), which although generally included in the Dasyuridce, should perhaps form the type of a family by itself; this animal differing from all the fore- going in the number, and also in the structure of the cheek-teeth, and thereby making a marked approximation to certain Marsupials of the Jurassic epoch noticed in the sequel. Before leaving this family, it should be mentioned that certain extinct Marsupials from the Tertiaries of Patagonia, referred to in the next chapter, seem to be inseparable from it, while there are strong reasons for regarding one of them (Prothylacinus) as very nearly allied to the existing genus Thylacinus. The last of the Australian families {Notary ctidce] of the sub- order is represented solely by the marsupial mole (Notoryctes), from the sandy deserts of central South Australia ; this being the only member of the order which has taken to a subterranean mode of life. There are no extinct Australian genera of the sub-order. Exclusive of the bats, the only other order of mammals well represented in the Australian region is that of the Rodentia, or Gnawing Mammals, which bear, how- ever, a small proportion to the marsupials, and all of which belong to the mouse-family (Murtdce). And it is noteworthy that although several of these belong to generic types unknown elsewhere, the whole of them are animals of comparatively small size, so that it is possible that their ancestors may have been introduced without a direct land-connection with any other part of the world. A curious feature in connection with this group is that two of the Australian species, namely Hydromys chrysogaster and Musfuscipes, are aquatic in their habits ; whereas, as we have seen, none of the Australian marsupials are natatorial, although the duckbill is eminently so. The Australian water-rat (Hydromys), which is common to Australia and New Guinea, belongs to a sub-family typically distinguished from all other Murtdce by the reduction of the molar teeth to two pairs in each jaw. While this animal has partially webbed toes, and is strictly aquatic in its habits, the allied Xeromys from Queensland is terrestrial, and approximates to the more typical members of the family, although to which group is still uncertain. The only other representatives of the sub-family Hydromyina are met with in the mountains of Luzon, II.] RODENTS. 41 in the Philippine group, where there is one genus allied to Hydromys, while other species have been assigned to the genus Xeromys. Whereas the typical Australian representative of the latter has but two pairs of molar teeth, one of the Philippine forms has three, thus approximating to more ordinary murines. The occurrence of these Australian types of rats in the Philippines is of the utmost importance in respect to Australia having received its mammalian fauna from south-eastern Asia. Of the typical genus Afus, whose geographical distributional area includes the whole of the eastern hemisphere with the ex- ception of Madagascar and many of the Polynesian islands1, Australia has upwards of twenty-six representatives, while two species occur in the Duke of York group, and others probably on the Papuan mainland. One of the Duke of York species (M. prcetor) ranges eastwards into the Solomons, where three other kinds are also found. The jerboa-rats (Conilurus*) form a pe- culiar saltatorial group restricted to the sandy deserts of the mainland of Australia, where they are represented by about a dozen species ; while the broad-toothed rat (Mastacomys) is con- fined to Tasmania, although its fossilised remains, like those of the other genus, are met with in the caverns of New South Wales. More nearly allied to the true rats and mice, the mosaic-tailed rats (Uromys) inhabit Queensland and the Aru Islands, one of the species from the former area also occurring in the Solomons. Lastly, the prehensile-tailed rat (Chiruromys], from the mountains of south-eastern New Guinea, represents a genus distinguished from all other placental mammals of the eastern hemisphere, with the exception of the British harvest-mouse and the Oriental binturong, by the prehensile nature of its tail. In connection with these rodents it is important to observe that fossil Murida are unknown from any part of the world earlier than the Miocene epoch, so that it is evident the living Australian representatives of the family are comparatively recent immigrants into the region they inhabit. 1 The Pacific rat (Mus exulans) appears to be widely distributed in these islands, see note on p. 29. 2 Commonly known by the preoccupied name Hapalotis. 42 THE NOTOG^IC REALM. [CHAP. Much discussion has taken place with regard to the date of introduction of the native dog, or dingo ( Canis dingo] Carnivores. . ,. . . -jj^i^-j. into Australia, and it was long considered that it was imported by human agency. Seeing, however, that its remains have been found in association with those of extinct kangaroos and Diprotodon, there seems considerable probability of its being an indigenous inhabitant of the country l. The only other non-volant mammal found in the Australian region is a species of pig (Sus papuensis). This animal is, however, so closely allied to certain Malayan species that it seems quite possible that its introduction may be due to human agency. The Australian region contains representatives of all the families of Bats with the exception of the Neogaeic Phyllo- stomatid(z\ some of the genera, such as the tube- nosed bats (Uronycteris*}, among the Pteropodida, being peculiar to this and the Austro- Malayan region, while others are more or less widely spread, or even cosmopolitan. It will be unnecessary to mention the various genera by name ; but the affinity of the Notogaeic Chiroptera to those of Eastern Arctogaea, as exemplified by the abundance of fruit-bats (Pteropodidce) and the absence of the PhyllostomatidcR) is noteworthy. In the following synoptical list the higher groups and genera (exclusive of Bats) peculiar to the Notogaeic realm are printed in italic type ; the letters A, P, and M Papuan following the names respectively indicate that the Genera. groups in question occur in Australia (inclusive of Tasmania), New Guinea (with the adjacent Papuan islands), or the Austro-Malayan region ; extinct groups have an asterisk prefixed. I. MONOTREMATA. Ornithorhynchidce, A. OrnithorhynchuS) A. EchidnidcB, A. P. Echidna, A. P. (The living species common to both areas.) Proechidna, P. 1 See Ogilby, "Catalogue of Australian Mammals," Sydney, 1891 — 92. 2 This name replaces the preoccupied Harpvia. II.] LIST OF MAMMALS. 43 II. Marsupialia. i. DlPROTODONTIA, A. P. M. Macropodidce, A. P. Mac r opus, A. P. Petrogale, A. Onychogale, A. Lagorchestes, A. Dorcopsis, P. Dendrolagus, A. P. Lagostrophus, A. Potorous, A. Caloprymnus, A. Bettongia, A. sEpyprymnus, A. Hypsiprymnodon, A. *Palorchestes, A. * Procoptodon, A. *Sthenurus, A. *Triclis, A. Phalangerida, A. P. M. Koala, A. Phalanger, A. P. M. Trichosurus, A. Pseudochirus, A. P. Petauroides, A. Dactylopsila, A. P. Petaurus, A. P. M. Gymnobelideus, A. Dromicia, A. P. Distcechurus, P. Acrobates, A. P. Tar sip es, A. *Thylacoleo, A. *Diprotodontidcz, A. * Diprotodon, A. (Elsewhere represented only by an aberrant group in South America.) 44 THE NOTOG^IC REALM. [CHAP. 4. *Nototheriid(K) A. * Nototherium^ A. 5. Phascolomyid&i A. PhascolomyS) A. *Phascolonus, A. ii. POLYPROTODONTIA, A. P. 1. Peramelid&i A. P. Peragale, A. Perameles, A. P. Chc&ropuS) A. 2. DASYURID^E, A. P. ThyladnuS) A. SarcophiluS) A. Dasyurus, A. P. Phascologale, A. P. Smmthopsis, A. AntechinomyS) A. MyrmecobiuSj A. 3. Notoryctidce, A. Notoryctes, A. III. Rodentia. i. MuRiDjE. Cosmopolitan. Hydromys, A. P. Xeromys, A. and Philippines. Mus, A. P. M. Conilurus, A. Mastacomys, A. Uromys, A. P. Chiruromys, P. IV. Carnivora. CANID^:, Cosmopolitan. Canis, A. Cosmopolitan. V. Ungulata. SUID.E. Throughout Eastern Hemisphere, except Australia. Sus, P. Elsewhere throughout greater part of Eastern Hemisphere. II.] ITS REGIONS. 45 VI. Chiroptera. All the families, with the exception of the Neogaeic Phyllostomatidae, well re- presented. The Polynesian region, as already said, is characterised by the general absence of non-flying mammals, and there- fore claims but little notice here. The only mar- supial occurring within the region is a variety of the widely-spread grey cuscus (Phalanger orientalis], which occurs in the Solomon Islands, where four species of Mus are likewise met with. As the cuscus, together with one of the rats, is also found in the Duke of York group, it may be inferred that the non-volant mammals of the Solomons have been derived from the latter area. In addition to members of widely-spread types, the Solomons possess two peculiar genera of bats. New Zealand appears to be inhabited only by two peculiar generic types of bats, each represented by a single species, and a rat (Mus maorimri}, but whether the latter is indigenous or intro- duced appears doubtful. Although a work devoted to mammals has little to do with an area where the sole member of the class is a bat of the genus Atalapha, brief mention must be made of R^on*"' the Sandwich Islands, which from their bird-fauna are regarded as entitled to distinction from the Polynesian region. Of the birds of this area, Mr W. L. Sclater ' writes that the greater number not only of the species, but even of the genera "are peculiar and wholly restricted to these islands. It is, of course, among the smaller land-birds (Passeres) that this individuality is most marked; but even in the other groups, where the distribu- tion is generally wider, the Hawaiian birds are, in many cases, local." Poverty, and an admixture of Australian and Malayan types, with a very marked preponderance of the latter, are the leading features in the mammalian fauna of the lay^n Region. Austro-Malayan region. This area includes the islands lying between Makassar Strait and the narrow channel separating Lombok from Bali on the west and the Australian region 1 Appendix, No. 28. 46 THE NOTOG^IC REALM. [CHAP. on the east. The largest of these is Celebes, while those of the Moluccan group, such as Gilolo, Buru, Ceram, and Timor-Laut, together with Timor and Sumbawa, are of smaller size. Unfor- tunately no complete lists of the fauna of these islands, so far as I am aware, have yet been published. Commencing with Timor and the Moluccas, we find several of the latter group of islands inhabited by four species of cuscus (Phalanger\ two of which are common to the Australian region, while the third (P. ornatus] is peculiar, and the fourth (P. celebensis], which in this group is found only in Sanghir Island, is an inhabitant of Celebes, where the other three are unknown. The only other Austro-Malayan marsupial1 is a variety of the Australian lesser flying-phalanger (Petaurus breviceps\ this variety ranging eastwards from Gilolo to the New Britain group. With the possible exception of certain shrews, most of the few Moluccan species of eutherian mammals appear to be identical with those of Celebes, whence they were probably introduced. A deer from Timor has received a distinct name (Cervus timoriensis\ and the same island is also inhabited by a common Malayan monkey (Macacus cynomolgus), a palm-civet (Paradoxurus hermaphroditus], and a true civet ( Viverra tangalungci], the latter being common to the Moluccas. The common Javan porcupine (Hystrix javanica), which is widely spread in the Malayan islands, is also found in Timor. There is likewise a cat in the same island, which although described under the name of Felis megalotis as a distinct species, and regarded by Mr Jentink as such, has been identified by Mr W. L. Sclater in his " Catalogue of the Mammalia in the Indian Museum " as a mere variety of the domestic species. In regard to all the Timorese forms which are closely allied to, or identical with well-known Malayan species, it is necessary to take into consideration the well-known partiality of the Malays for taming wild animals and carrying them about during their voyages; and it is highly probable that all or most of such animals found in Timor have been thus introduced. From the small island of Flores Mr Jentink has described a rat (Mus armandvillei), which is the largest member of its genus. 1 The Kei Islands, like the Aru group, may be best affiliated to Papua. II.] AUSTROMALAYAN REGION. 47 In addition to the above-mentioned cuscus, which appears to be its only marsupial, Celebes possesses several peculiar types of eutherian mammals. Among these is a black and nearly tailless ape (Cynopithecus niger) representing a genus by itself; while there is also a species of macaque (Macacus maurus) peculiar to the southern portion of the island. A species of the lernuroid tarsiers (Tarsius fuscomanus) is found both in Celebes and the neighbour- ing islands of Salayer and Sanghir, although represented by an FIG. 5. FORE PART OF SKULL OF BABIRUSA (Bcibirusa alfurus], allied form in the Philippine group. In the Carnivora there is a Malayan species of civet ( Viverra tangalunga), and also a peculiar species of palm-civet (Paradoxurus musschenbroecki). In the pig- tribe the babirusa (Babirusa alfurus), characterised by the extra- ordinary development of its tusks, is the sole representative of a genus confined to this island and Bum; while scarcely less peculiar is the small and somewhat antelope-like buffalo known as the anoa (Bos depressicornis\ which although allied to the tamarao (B. mindo- 48 THE NOTOG^IC REALM. [CHAP. rensts)1 of the Philippines, has its nearest relatives in certain extinct species from the Pliocene of Northern India. There is also a true pig (Sus celebensis], nearly allied to Malayan forms ; as well as a deer forming a variety of the widely spread sambar (Cervus unicolor}. Among the rodents, a rat with an extremely long muzzle constitutes a peculiar genus (Echinothrix)^ and there are also other Muridce, as well as squirrels (Sciurid 60 THE NOTOG^IC REALM. [CHAP. that the giant flightless species such as Phororhachis and Bron- tornis of the Patagonian Tertiaries have been recently shown to form a totally distinct group — the Stereornithes, — and it is quite probable that the same may prove to be the case with Gastornis, Dasornis, and Diatryma of the lower Eocene of the northern hemisphere. Apart from these, the earliest known ratites are Hypselornis of the Pliocene of India — which appears to be allied to the emeus and cassowaries — and the Australian Dromornis, one species of which is likewise of Pliocene age; all the other forms being Plisiocene. Moreover, it is now tolerably certain that the true ratites have originated from flying birds, and it is therefore highly probable that the group is an essentially modern one1. Accordingly, there is a strong presumption that the ances- tors of these birds did not enter Notogsea till comparatively late in the Tertiary period ; and that, in fact, their southern migration was not far removed in time from that of the giant land-tortoises, noticed in the next chapter. Possibly they may have entered Australia by way of New Guinea during the connection which Mr Hedley believes to have existed between those two countries late in the Tertiary epoch ; while the New Zealand forms may have made their way by means of the presumed land-connection between these islands and the Solomons, New Hebrides, etc. Of course there is the difficulty as to why mammals did not enter the Polynesian region at the same time; but it is conceivable that even at this date the mammalian fauna of South-eastern Asia may have been very poor in Eutherians, while it is quite possible that the ancestral forms of these birds may not have required the complete land-connection necessary for the passage of the higher mammals. Such connection as served for these birds, however, may have well sufficed for the transit of the ancestors of the Australian murine rodents, which almost certainly entered the country at a later date than the original marsupials and mono- tremes. Assuming, then, that the marsupials and monotremes of the Australian region did not reach their present home till the early part of the Tertiary epoch, we must make the further assumption 1 Captain Hutton is of opinion that the moas originated directly from flying birds in New Zealand, but the evidence in favour of this view appears insufficient. II.] ORIGIN OF MONOTREMES. 6l that at this period South-eastern Asia was entirely, or to a great extent, devoid of higher mammals. Nor is this unlikely, seeing that the ungulates and carnivores of the lower Eocene of the northern hemisphere would clearly have required time to spread themselves to the southward. Hence it may be suggested that, towards the close of the Cretaceous epoch there was first a migra- tion towards the south-east of the ancestral marsupials (and monotremes) inhabiting the northern hemisphere during the Secondary epoch ; and that similar migrations of the higher mam- mals took place during Tertiary times. When once the ancestral polyprotodont marsupials obtained a footing in New Guinea and Australia, where they have since been isolated from any serious competition with the higher mammals, they flourished and developed to a degree which they could not possibly have attained in any other part of the world under existing conditions. And it is doubtless within this region that the more specialised diprotodont types were evolved. Remark- able as it undoubtedly is, the present state of development of the Australian marsupials is nothing to what it was during the Plistocene period, when there lived the giant kangaroos, pha- langers, wombats, diprotodons, and nototheres already alluded to, by the side of which the largest existing species would appear almost dwarfs. The cause of this universal extinction (for uni- versal it is) of all the larger types of mammalian life throughout the world soon after the appearance of man, is one of those problems which at present is not capable of being satisfactorily solved, as not even a glacial period could have made a clean sweep of the whole globe. It may be added that the evolution of the diprotodont marsupials within the limits of the Australian region, points to the conclusion that the outlying discuses of the Austro-Malayan regions are immigrants from the south-east. With regard to the monotremes, it has already been mentioned that there is no record of their past history beyond the limits of the Australian region. It can, however, scarcely be doubted that their ancestors came from the north with the primitive marsupials ; and if, as is not improbable, the Secondary and early Tertiary Multituberculata of the northern hemisphere are an allied type, there can be no doubt whatever as to this having been the case. 62 THE NOTOG^IC REALM. [CHAP. That Notogaea, as typified by the Australian region, is entitled to form one of the three primary zoological divisions of the globe, the distinctness of its mammalian fauna from that of any other area, not only at the present day, but likewise during the Plisto- cene, and probably also the Pliocene epoch, amply demonstrates. The inclusion within the same realm of the Polynesian region, — which evidently never had such a close connection with south- eastern Asia during the time that area was mainly populated with marsupials, — is justified partly on account of its containing more or less similar types of birds, and partly by the practical absence of terrestrial mammals. On the other hand, the Austro-Malayan region, which is really a kind of zoological No-man's-land, is placed within the limits of the same great realm more as a matter of convenience than anything else, although it is undoubtedly sharply differentiated from the Oriental region by Wallace's line. In conclusion, a few lines may be devoted to showing that certain other groups indicate that the vertebrate fauna of Notogaea, as a whole, has had a northern origin. Among the lizards, the family of iguanas (IguanidcB], which in this realm occurs only in the Fiji and Friendly Islands, is represented in a fossil state in the Oligocene beds of France ; while the gigantic extinct monitor ( Varanus priscus) of the Australian Plistocene appears to have its nearest ally in the smaller V. sivalensis of the Pliocene of northern India. The Notogseic Chelonians, which are confined to Australia and New Guinea, all belong to the side-necked group (Pleurodira) of the order, and are represented by the families Chelyidce. and Carettochelyidce, the latter containing only a single species from the Fly River. Now, although none of the Australian genera have been detected in the northern hemisphere, the side-necked chelo- nians, as shown in the next chapter, were abundantly represented there during the early Tertiary and Secondary epochs ; and it is a remarkable fact that an extinct genus believed to be allied to Carettochelys occurs in the Eocene of northern India. Although from their aquatic, and sometimes partially marine habits, the crocodiles are of less importance than some other groups from a distributional point of view, yet it is noteworthy that the single representative of that group ( Crocodilus porosus] inhabiting Noto- gaea (where it is found in North Australia, the Solomons, and II.] SURVIVAL OF OLD FORMS. 63 Fiji) is spread over India, Ceylon, and the south of China ; while the absence of caimans and jacaras from Notogaea aifords, so far as it goes, an additional argument that any land connection which may have existed in Tertiary times between Australia and South America must either have been very transitory, or must have been situated in such latitudes that tropical forms could not have used it as a means of transit. Of more importance than all is the tuatera lizard (Sphenodon) of New Zealand — the sole existing representative of the order Rhynchocephalia, — since this curious creature is closely allied to the extinct Rhynchosanrus and Hyperodapedon of Triassic strata of the northern portion of the Old World. Finally, the Port Jackson shark (Cestracion philippi) belongs to a genus which was living in the seas of Europe during the Jurassic and Cretaceous periods ; and the sole living survivor of the. swarms of species of the Molluscan genus Trigonia inhabiting the same seas occurs in Australian waters. CHAPTER III. THE NEOG^IC REALM. Extent and Characters — Mammaliferous Deposits — Monkeys — Bats — Insecti- vora — Carnivores — Ungulates — Horses — Litopterna — Astrapotheria — Toxodonts — Pyrotheria — Proboscideans — Rodents — Edentates — Arma- dillos and Glyptodonts — Sloths — Anteaters- — Ground-sloths — Marsupials — Cetaceans — Early Distinction of the Neogaeic Fauna — Early Separation of N. and S. America — Incursion of Northern Mammals — Distinctness of the existing Fauna — Origin of the Santa Cruz Fauna — Antarctica and the South American element in the Ethiopean Fauna — Conclusion — Sub- regions. THE second primary zoological division of the globe may be known as Neogsea1, or the Neogseic realm. It characters"** includes only the Neotropical region. Comprising not only the whole of South and Central America, as well as the West Indian Islands, this area also embraces the lowlands lying on either side of the Mexican plateau — the so- called tierras calientes — thus running up in a fork-like manner to the lower extremity of North America. While, therefore, the greater part of this vast area is sharply delineated by its coast- boundary, to the north it has a kind of No-man's-land connecting it with the Sonoran region of Arctogaea, and, as will be shown later, through this transitional area there has been a certain amount of intermixture of the proper faunas of the Neotropical and Sonoran regions. Neogaea, as a whole, may be characterised as a country of extensive tropical forests or open grassy plains ; deserts occupying only a few scattered areas in the upper Argen- tine (Tucuman, etc.), and certain parts of the coasts of Chili 1 This term was originally proposed by Dr Sclater to include the whole of the New World, but has been used by an anonymous writer (Appendix, No. 4) in the present sense. Dr Sclater 's term Dendroggea (Appendix, No. 27, p. 214) is open to considerable objection, as the greater part of Argentina is woodless. CHAP. III.] EXTENT AND CHARACTERS. 6$ and Peru ; the whole of the rest of the area, with the exception of the higher regions of the Andes, being thus admirably adapted for the support of animal life. At least one half of the whole area is occupied by a dense tropical forest, attaining its richest develop- ment in the hot steamy tracts of Brazil and Paraguay, and being unequalled in extent in any other part of the globe. With a width of some three thousand miles from the Atlantic seaboard at Per- nambuco to the foot of the Andes, this forest extends north and south for nearly thirty degrees of latitude ; while not only does it clothe the lowlands and valleys, but extends high up the mountain- sides, as may be seen in the exquisitely lovely harbour of Rio de Janeiro, where the forest-vegetation commences immediately above the wash of the waves, and thence extends in one continuous leafy mass to the summits of mountain-ranges at an elevation of eight or nine thousand feet. In the northern part of the area open grass-lands, like the "campos" of Brazil and the savannas of Venezuela, alternate with the forest ; while in the neighbourhood of Buenos Aires the open pampas1 forms one extensive sea of grass. The Andes, constituting the backbone of the country, run in one continuous chain from north to south on the Pacific seaboard, and present the usual varieties of climate and physical conditions common to other elevated mountain-ranges. Such climatic variations are, however, only an epitome of those met with in travelling from the northern to the southern extremity of the area ; the steamy valley of the Amazons having a tropical climate, whereas when we reach the southern point of Patagonia and Tierra del Fuego we are in the midst of snows and glaciers. To the hot forest-regions are restricted the monkeys, marmosets, sloths, ant- eaters, and tree-porcupines ; while the open plains of the south are tenanted by guanaco, deer, viscachas, and rheas. Moreover, in the forests, the variety of mammalian life, especially as regards the larger forms, is in marked contrast to its comparative paucity in the open plains ; not but that, till civilised man made his appear- ance on the scene, the number of individuals may have been nearly, if not quite as large in the latter area as in an equal extent of the former. 1 Although in Spanish the term 'pampas' is plural, in English it seems preferable to use it as singular. L. 5 66 THE NEOG^IC REALM. [CHAP. At the present day the mammalian fauna of Neogaea is mark- edly distinct not only from that of Notogaea but likewise from that of the whole of the rest of the globe (Arctogsea), although the dis- tinction is now, owing to free communication with the north, much less marked than it was in Tertiary times, and it is accord- ingly essential to enter at once into the consideration of the extinct forms in order to show why this part of the world is entitled to rank as one of three primary zoological regions. There are several districts in South America where fossil Mammal- remains of mammals have been found ; most of these iferous being remarkable for the extraordinary profusion in which the bones occur. The first that may be men- tioned are the celebrated caves of Lagoa Santa, in the province of Minas Geraes, to the northward of Rio, which have yielded re- mains of a great variety of Plistocene genera and species, inclusive of those of man. Probably contemporaneous with these are the sand-dunes on the coast of Buenos Aires, which likewise contain human remains in association with those of extinct mammals ; while the so-called Pampean beds of the Argentine pampas are apparently somewhat older, although still pertaining to the Plisto- cene period. As these Pampean deposits are exceedingly rich in fossil mammals, they may be described in some detail. They form the great level tract of country extending southwards from the Rio de la Plata and the Parana to the Rio Colorado, south of Bahia Blanca, and westwards from the Atlantic seaboard about half the distance to the Andes; thus occupying some 200,000 square miles of country. The pampas is an almost level grass- covered plain, intersected by water-courses, and penetrated near its margins by small mountain-ranges, while it is almost entirely barren of trees. It is composed of a- rich black alluvial mud, mingled with beds of sand, and underlain by, or in some places interstratified with layers of a hard white calcareous deposit known as tosca\ but in certain spots it contains beds of marine shells belonging to species still living in the adjacent seas. Except in those spots where the tosca comes to the surface, there is not a stone or a pebble to be seen in the whole deposit, and near Buenos Aires the formation has been bored to a depth of ninety feet without touching bottom. From its composition it is evident III.] MAMMALIFEROUS DEPOSITS. 67 that the deposit has been carried down from the interior of the north by the Parana, Paraguay, and other tributaries of what is now the Rio de la Plata ; but since there is no splitting of the latter river at its estuary, it is evident that the formation cannot properly be called a delta. That it is mainly of freshwater origin seems evident not only from its intrinsic character, but likewise from the vast number of entire skeletons of mammals buried within it, since these creatures must certainly have lived very near to the places where their bones are now entombed. In the more southern part of its area the pampas is, however, probably to a large extent of estuarine origin ; and the presence of layers of marine shells in its uppermost horizon near Buenos Aires proves that at least a portion was submerged beneath the sea before its final upheaval. Whereas the Rio de la Plata now flows in a single channel in a south-easterly direction near the northern limit of the coast-portion of the pampas, it would seem probable that the Parana and Paraguay rivers may have originally continued their southerly course across the southern pampas, through which they may have flowed in a number of streams. Most likely the Pam- pean formation was laid down in a slowly subsiding area, in which the rate of deposition approximately counterbalanced the sinking, so that the greater part of it has been always land until the period of the great submergence. After the latter, the entire area was upheaved to a small degree above the sea-level, when the rivers assumed their present approximate courses. Whereas in certain localities the deposit is barren of mammalian remains, in other spots it appears absolutely crowded with them, and the number of entire skeletons that are entombed in it must doubtless be counted by thousands. Somewhat older than the Pampean is a mammaliferous deposit occurring on the coast near Bahia Blanca in a small hill known as Monte Hermoso ; and beds of approximately equivalent age occur in Catamarca at the foot of the Andes. Probably these beds should be regarded as of Pliocene age ; and it may be mentioned that equivalents of these deposits are met with in other parts of Argentina, while representatives of the Pampean occur in Pata- gonia, Chili, Bolivia, etc. Still older than the Monte Hermoso deposits are the Santa Cruz beds of Patagonia, occurring not only 5—2 68 THE NEOG^IC REALM. [CHAP. on the river of that name, but likewise further north in the Chubat district. These Santa Cruz beds are exceedingly rich in mamma- lian remains, which are stained of a deep black colour ; and while they contain many groups common to the Pampean formation, they lack those forms found in the latter which have an Arctogseic fades, thus indicating that we have reached an epoch when the fauna of South America was far more completely isolated from that of the rest of the world than is at present the case. By Dr Ameghino the Santa Cruz beds were at first correlated with the lower Eocene of Europe, although he subsequently admitted that they must occupy a somewhat higher position in that period1. From the nature of their entombed mammals2 it is, however, certain that they must be still newer, and they cannot be regarded as older than lower Miocene. Indeed, they are underlain by the so-called Patagonian beds3, which are of marine origin, and contain numerous cetaceans, among which are whalebone whales (Mystacoceti). From equivalent beds in the Chubat district of Patagonia, a large number of such cetaceans have been described by the present writer4, and it is quite evident that the oldest age that can be assigned to these beds is upper Oligocene, seeing that in Europe whalebone whales are unknown till a considerably later epoch. Probably the freshwater beds containing the peculiar mammal alluded to below under the name of Pyrotherium are the freshwater equivalents of the Patagonian horizon5. Additional evidence in support of the comparatively late age of the Patagonian beds is afforded by the researches of Prof. Cope6 on the cetaceans of the Miocene (or Upper Oligocene?) of the United States. From these beds have been obtained remains of Hypocetus (Paracetus) — a genus elsewhere known only from the Patagonian beds — together with Cetotherium (also occurring in the latter deposits), Balcenoptera, and a species of Balcena identified with one from the European Pliocene ; the North and South American species of Hypocetus being closely allied. 1 Bol. Ac, Cordoba, Vol. xin. p. 260 (1894). 2 Remains of the existing genus Dasypus occur in these beds. 3 Ameghino, loc. cit. 4 Ann. Mus. La Plata,— Pal. Argent. Pt. II. (1893). 5 Ameghino, op. cit. p. 262. 6 Prof. Amcr. Phil. Soc. Vol. XXXIV. pp. 135 — 155 (1895). III.] MONKEYS. 69 With these preliminary considerations, we pass to a critical examination of the recent and fossil mammalian fauna of Neogaea, which will show how intimately investigations of this nature are connected with the present and past configuration of the land- areas of the globe. Like as are many of them superficially to their cousins of the Old World, the monkeys of the New World, which ,, , , .... . - Monkeys. are now confined to the tropical forest-regions of the Neogaeic realm, are structurally quite different to the former ; and this circumstance, coupled with their isolated distribution, in- dicates that their relationship is, at most, but a very distant one. Indeed it is not improbable that the Old and New World monkeys may have originated independently from the group of lemurs, which were formerly very widely distributed over Arctogaea. Of the two families of Neotropical monkeys, the first is represented by the beautiful little marmosets, constituting the family Hapalida. Although having the same number (32) of teeth as the Old World monkeys, the marmosets differ in that the number of premolars on each side of both jaws is three, and that of the molars two ; these numbers being transposed in the other group. The marmosets are further distinguished by the broad septum between the nostrils, the absence both of pouches in the cheeks and of callosities on the buttocks, and the want of any prehensile power in the tail ; in addition to which it may be noticed that the thumb is not op- posable to the other fingers, while all the digits, with the exception of the first on the hind foot, are provided with long claws. None of these animals are known from the Santa Cruz FlG. 9. PART OF RIGHT LOWER JAW OF HoniUHCulus. deposits. In the second family or Cebidce, which includes much larger species than the diminutive marmosets, the total number of 70 THE NEOG^EIC REALM. [CHAP. teeth is 36, owing to the retention of the three pairs of molars found in the Old World group ; these monkeys being further dis- tinguished from the marmosets by the presence of nails on all the fingers, and by the tail being frequently prehensile. The Santa Cruz beds have yielded remains of monkeys (JFTomunculus) refer- able to this family, showing that they belong to the original South American fauna, but beyond this nothing is known as to the palseontological history or origin of the group. Lemurs, both in the past and the present, are quite unknown in the realm under consideration. Although bats might be thought of comparatively little im- portance from a distributional point of view, yet the Bats. :_ Neogseic realm presents some very remarkable features in this respect. In the first place the two Old World families of fruit-bats (Pteropodtdce) and horse-shoe bats (Rhinolo- phidce] are entirely wanting, whereas the great family of vampire- bats (Phyllostomatida) is mainly restricted to it, although a few representatives straggle northwards along the Pacific coast of North America. The family of Emballonuridtz is also more strongly represented here than in any other part of the world. Probably on account of their small size, there is no palaeonto- logical record of the South American bats from the earlier Tertiary deposits. The Insectivora are almost unrepresented in the continental portion of the realm, although a shrew (Sorex) Insectivora. ' ° . reaches Guatemala and Costa Rica, and a member of the allied N. American genus Blarina is also found in the last-named country; both these being doubtless very recent immigrants from the north. Very remarkable is the occurrence in the West Indian Islands of the two species of Solenodon, consti- tuting a distinct family (Solenodontidce) by themselves. These have generally been considered as very nearly allied to the tenrecs (Centetidce) of Madagascar, but Mr O. Thomas1 is of opinion that the relationship is not really very close, the similarity in the structure of their molars being merely a generalised character. Both, however, probably indicate an ancient group, which has 1 Proc. Zool. Soc. 1892, p. 500. III.] CARNIVORA. 71 migrated from the higher latitudes of the northern hemisphere to find a refuge in these two far distant localities. It may be mentioned that the Solenodontidce and Centetidce, together with the Potamogalida of western, and the Chrysochloridcz, or golden moles, of southern Africa, constitute a section of the order distinguished from all the other forms by having the cusps on their upper molar teeth arranged in the form of the letter V, instead of in that of a W ; and it will not fail to be noticed that the whole group is now exclusively a southern one. Whether it is represented in the Santa Cruz beds is not quite certain, although one lower jaw has been referred to it by Dr Ameghino1 under the name of Necrolestes. Be this as it may, it is clear that, with the exception of the afore- said shrews, insectivores with W-shaped molars are entirely want- ing in the realm ; and that such V-shaped types as still exist, or formerly occurred, evidently indicate an ancient northern group. Other instances of the survival in South America and Madagascar of allied forms will be noticed in the sequel, and admit of a somewhat similar explanation. Although fairly well represented at the present day, the carni- vores of this realm have but few absolutely charac- . Carnivora. tenstic forms, and as no remains of the true Carni- vora occur in the Santa Cruz beds, the whole of them may be regarded as comparatively late immigrants from the north. The civet family ( Vsvtrrida) and hyaenas (Hyanidcz) are totally wanting at all epochs, as indeed they are throughout the whole of the New World; but the weasel tribe (Mustelidce) have a comparatively small number of representatives. Cats (Felidcz) on the other hand are numerous, although several species, and more especially the puma (Felts concolor), range to a greater or less extent into North America. Such a cosmopolitan genus is, however, of no importance whatever from a distributional point of view; and much the same may be said of the extinct sabre-tooths (Machar- odus), which were distributed in Tertiary times throughout the entire northern hemisphere. Unknown in the deposits of Monte Hermoso and Santa Cruz, this genus is represented by a gigantic species in the Pampean, which undoubtedly reached the country t 1 Bol. Ac. Cordoba, Vol. xm. p. 364 (1894). 72 THE NEOG^ilC REALM. [CHAP. from the north in company with the other late immigrants. The dog tribe (Canidcs) likewise comes under the category of cosmo- politan groups, and has numerous South American species, although only one from the Monte Hermoso beds dates earlier than the Pampean. It may be mentioned that while true wolves1 are absent from this realm, all the continental living members of the genus Cants found in it form a group by themselves, quite distinct from the true foxes of other regions. Very remarkable is the occurrence in the Pampean of a large species (C. moreni) perfectly distinct from all those now inhabiting the country, and presenting some curious approximations in the structure of the skull to domestic dogs. Peculiar to the realm is the bush-dog (Icticyori), of Guiana and Brazil, — a small short-haired and short-legged species differing from all other members of the family by the small size and reduced number of the molar teeth. Its remains occur in the Brazilian caves, but are unknown from earlier deposits ; and it may thus be regarded as a comparatively late immigrant from the north, which perhaps developed its special characters after its arrival in the country. Of the Ursidce. there is but a single existing species in Neogaea, namely the spectacled bear ( Ursus ornatus) of the highlands of Chili and Peru ; but there occur in the Pampean formation of the Argentine remains of an allied extinct genus known as Arctotherium, another species of the same genus being recorded from the superficial deposits of California. Far more characteristic of the realm are the raccoons and coatis (Procyonidf which range south of Paraguay. The extinct Castoroididcz include large beaver-like rodents with complex molars like those of the viscacha, and are represented by the typical Castoroides from the Plistocene of the United States, and likewise by Amblyrhiza (Loxomylus) from caves in the West Indies, and, it is said, also from the later Tertiaries of Argentina. This family is accordingly a northern type. The second or murine group of rodents (Myomorpha) contains five families, namely the dormice (Myoxidce), the jumping-mice and jerboas (Dipodidce), the mice and rats (Muridce), the mole-rats (Spalarida), and the American pouched rats (Geomytda). Out of all these, practically the only one represented in the realm is the cosmopolitan Muridce, although among the Geomyidce the genus Heteromys enters the transitional Mexican sub-region. In the 88 THE NEOG^IC REALM. [CHAP. Muridce, however, as in North America, the true rats and mice (Mus) of the Old World are entirely wanting, except through human introduction ; their place being taken by the white-footed mice (Sttomys) common to the whole of the New World, and nearly related to the European hamsters. Being essentially a northern type, they are certainly recent immigrants into South America from the north ; and the same is doubtless true of certain genera of the family now peculiar to this realm, such as the fish- eating rats (Ichthyomys\ of Peru and Venezuela, the groove-toothed mice (Rhithrodori), one of which ranges as far south as Tierra del Fuego, and the Brazilian genus Holochilus, which is another form allied to the hamsters. Although the cotton-rat (Sigmodon) occurs in Central America, and occasionally wanders still further south, the whole tribe of voles and their kindred are absent from the entire realm. Represented by six families, all of which occur within its limits, the porcupine-like group, or Hystricomorpha, may be regarded as the characteristic rodents of Neogsea. From both the preceding sections of the order the members of this section may be readily distinguished by the angle (or hinder inferior projection) of the lower jaw taking its origin from a prominent ridge running along the side of the jaw, instead of from the inferior edge of the socket for the incisor teeth. Of the families confined to this realm, that of the cavies (Caviidcz], includes heavily-built rodents, with four front and three hind toes, rudimental or short tails, and the cheek-teeth divided by transverse folds of enamel into a number of thin parallel plates. The genera of this family include the true cavies (Cavia] — so well known through the domestic guinea-pig — all of which are small short-limbed forms ; the larger and taller Patagonian cavy (Dolichotis}; and the carpincho, or capivara (Hydrocha>rus\ which is the largest living member of the order, and is characterised by the large number of plates going to form the last molar tooth in each jaw. Although they do not appear to have been recorded from the Santa Cruz beds, remains of members of this family occur in the Parana1 horizon, and also in that of 1 There is some confusion with regard to the age of the Parana beds. They are overlain by marine strata which have been identified with the Patago- nian ; but it is more probable that they are newer than the Santa Cruz beds, III.] RODENTS. 89 Monte Hermoso. Of the former Plexochairus differs from Hydro- chorus only by the somewhat simpler structure of the last molar, Hydrochcerus itself occurring in the Monte Hermoso beds. Other forms from the Parana stage are Eucardiodon and Cardiotherium, apparently more nearly allied to Cavia. Here it should be mentioned that certain European Oligocene genera (Issiodoromys and Nesocerodori) have the molars so like those of the cavies that they are regarded by Dr Schlosser as nearly allied' to the family. By Prof. Zittel they are, however, included in the extinct family Theridomyidce, which is classed with the dormice and certain other families in a group regarded as intermediate between the Sciuromorpha and Hystricomorpha. If, as would seem probable, they are really allied to the latter, they are of the utmost import- ance as indicating a connection between the middle Tertiary rodent fauna of Europe with that of South America1. Nearly allied to the cavies are the agutis (Dasyprocta) and pacas (C&logenys), collectively constituting the family Dasy- proctidcE, and differing from the former in that the folds of enamel merely form notches on the sides of the crowns of the cheek-teeth. In a fossil state the family seems to be only known by remains of the existing genera from the Brazilian caves. The third peculiar family (Dinomyidce) is known merely by a single specimen from Peru. The only other family exclusively confined to the realm is that of the Lagostomatidcz (Chinchillidce], which includes not only the true chinchillas (Eriomys) and Cuvier's chinchilla (Lagidium), but likewise the viscacha (Lagostomus] of the Argentine pampas. All the members of this family have long bushy tails, elongated hind limbs, and the cheek-teeth divided by complete transverse folds of enamel into thin plates. Lagostomus occurs fossil not only in the Pampean, but likewise in the older Tertiaries of Parana; while it may be doubted whether Pliolagostomus and Prolagostomus of the Santa Cruz beds are really entitled to generic distinction. Other allied forms from the latter deposits are although their lowest portion is older than the Monte Hermoso stage. (See Ameghino, Bull. Ac. Cordoba, Vol. xiil. pp. 260, 261, 1894.) They may be partly made up of the remains of pre-existing beds. 1 Dr Schlosser writes me to the effect that he is fully assured these forms are the ancestors of the Caviidce. 90 THE NEOG-^IC REALM. [CHAP. described under the names of Perimys and Sphodromys, and thus indicate that the family is essentially South American. The largest known rodent is the extinct Megamys, from the Parana beds ; the typical species of the genus being described as equal in size to an ox. The porcupines (Hystricida), which form a practically cosmopolitan family, sufficiently distinguished by their spiny covering, are represented in South America by the two arboreal genera Synetheres and Chcztomys, differing from all their allies by their prehensile tails, although otherwise related to the North American Erethizon, with which they form a separate sub- family. Of Chcetomys an extinct representative occurs in the cavern-deposits of Brazil; and in the Santa Cruz beds the family is represented by apparently extinct generic types described under the names of Stiromys, Acaremys, and Sciamys. Since fossil remains of Hystrix date from the European Miocene or Oligocene, there is distinct evidence of a connection between the early South American rodents and those of the Old World ; while as Erethi- zon is first known from the Plistocene of North America, it may probably be regarded as a late immigrant into that country from the south. The largest of all the families of the Hystricomorpha is that of the Octodontidal, in which out of a total of nineteen genera upwards of fifteen belong to the realm under consideration, while the remaining four are African and mainly Ethiopian. In addition to other features, the rodents of this family have the crowns of the cheek-teeth marked by infoldings of enamel from both sides; there are usually five toes to each foot ; and the general form is more or less rat-like. Of the Neogaeic forms, the typical genus Octodon is represented by the degu of Chili and Peru, which is a large rat -like animal, with a brush-tipped tail; other species occurring in Bolivia. The latter country is likewise the home of the allied genus Habrocoma, the members of which vie with the chinchillas in the delicate softness of their fur. Nearly related are the burrowing South American tuco-tucos (Ctenomys\ characterised 1 By some the family is divided into three; viz. Capromyidce, with the West Indian Capromys, the S. American Myopotamus, and. the African Triaulacodus ; Ctenodactylidcz, including the remaining African forms; and Octodontidce, comprising all the other American types. III.] RODENTS. 91 by the comb-like bristles on the hind feet, and their bell-like cry ; two Chilian species, constituting the genus Spalacopus, being distinguished by their rudimental ears. Schizodon, on the other hand, of which there is a single species from the Southern Andes, has the ears larger than in the tuco-tucos. The South African Petromys has been regarded as very closely allied to Spalacopus^ but it is more probable it should be classed with the other two African genera Ctenodactylus and Pectinator, to constitute a separate sub-family. The third sub-family includes all the other genera, one of which (Triaulacodus1}, as represented by the cane- rats, is Ethiopian, while the whole of the remainder are Neogaeic. Many of the species are of large size, some being arboreal and others aquatic. Of the latter the best known and largest is the coypu (Myopotamus), widely spread over South America ; while in the West Indies the group is represented by the equally large arboreal hutias (Capromys and Plagiodori). The other seven genera are South American, and include smaller rat-like forms, which in the case of the two genera Loncheres and Echinomys have flattened spines mingled with the fur; the others being known as Mesomys, Dactylomys, Cannabateomys, Cercomys, and Cartero- don. Several of the existing genera occur fossil in the caves of Brazil and the Pampean ; Myopotamus also occurring in the infra-pampean beds of Parana, together with the reputedly extinct forms described as Orthomys and Morenia. Other extinct genera, such as Neoremys, Scleromys, and Addphomys occur in the Santa Cruz deposits, and appear to be very closely allied to Myopotamus. It is noteworthy that an extinct Octodont (Pelle- grinia) allied to Ctenodactylus occurs in the Plistocene or Pliocene of Sicily; while Ruscinomys from the Pliocene of France is believed to belong to the same group. Finally, the genus Eocardia, together with certain other allied forms from the Santa Cruz beds, are regarded as indicating a separate family (Eocar- ditdce) of the section. With regard to the extinct family Theridomyida from the middle and upper Oligocene of Europe, which includes not only Theridomys and Archceomys, but probably also the above-men- 1 The name Aulacodus being preoccupied, that of Triaulacodus is proposed in substitution. 92 THE NEOG^IC REALM. [CHAP. III. tioned Nesocerodon and Issiodoromys (p. 89), it seems highly probable that these are really the ancestral forms of the modern Hystricomorpha, although their lower jaws approximate to the general type of those of the more generalised Sciuromorpha and Myomorpha. The last section of the order (Lagomorpha), which includes the hares and picas, and is essentially a northern one, is but poorly represented in Neogaea ; the picas (Lagomyidcz) being unknown there, while in the whole of the realm there are only two species of Leporidce, one of which is Brazilian. It has been usual in zoological systems to include under the title of Edentata not only the armadillos, anteaters, and sloths of South America, but likewise the Old World pangolins (Manida) and aard-varks (Orycteropodidaz). There can, however, be no doubt that there is little or no connection between the two groups, and the latter may accord- ingly be separated as a distinct order under the title of ErTodientia. In this restricted sense the edentates at the present day are, per- haps, the most characteristic and remarkable of all the Neogaeic mammals. Whereas, however, the sloths and anteaters are entirely Neogseic, a few of the armadillos have wandered at a comparatively modern date as far north as Texas ; but this does not detract from their essentially southern character, seeing that they are well represented in the Santa Cruz beds, and, if we exclude certain remains of doubtful affinity from the Oligocene of France, are quite unknown elsewhere. This, however, is by no means all, since there are two extinct families of the order, dating from the Santa Cruz beds, which were extremely abundant during the Pliocene and Plistocene epochs ; some few of these having managed to obtain an entrance into North America about the Miocene epoch. Central and South America may accordingly be considered as essentially the home of the edentates ; and are thus broadly demarcated from all other parts of the world. It would be superfluous to point out all the distinctive features of the order, but it may be mentioned that none of the living forms have teeth in the front of the jaws ; while in all those genera in which teeth are present, these are of comparatively simple structure, being unprovided with a coating of enamel, growing 94 THE NEOG^IC REALM. [CHAP.III. continuously without ever forming roots, and being mostly very similar throughout the series. The mailed, or loricate edentates are represented by the two families of the armadillos (Dasypodidce] and glypto- Armadillos donts (Glyptodontida], the latter of which died out and Glypto - at the close of the Plistocene or the commence- ment of the Recent epoch, whereas the former is still abundant. With the exception of the two species of pichiciagos (Chlamydo- phorus] from Mendoza and Bolivia, in which a solid coat of mail is confined to the hinder region of the body, the members of both families have their bodies protected by a bony armour, while their heads are guarded above by a shield of the same nature, and their tails are enclosed in a tubular sheath. Covered exter- nally with horny shields, like the shell of a tortoise, the carapaces of these animals are formed of a number of small plates of bone, either united everywhere by their edges into a continuous solid armour, or in the middle region of the body overlapping one another like the tiles on a roof. In the true armadillos (of which the living Argentine forms are all comparatively small creatures, although one Brazilian species reaches nearly a yard in length) the carapace consists of a nearly solid buckler in front and behind, while between the two are situated a variable number of movable over- lapping bands, which in some instances admit of the body being rolled up into the form of a ball. They have all long snouts, and simple, subcylindrical teeth. The true existing armadillos may be divided into the genera Dasypus, Xenurus, Priodon, Tolypeutes, and Tatusia. The first of these, in which there are six or seven movable bands to the carapace, is found throughout the Argentine Tertiaries to the Santa Cruz beds, one of the fossil species from the higher beds of the series having a skull of nearly a foot in length, and thus vastly exceeding all its living congeners in size. Tatusia, in which the carapace has from seven to nine movable bands, does not appear to be known below the Pampean beds, where it is represented by the large species of which the external skeleton is shown in the figure on p. 93. A third genus, Eutatus, which likewise comprises species of large size, and ranges from the Pampean to the Santa Cruz beds, is distinguished by having over thirty movable bands in the carapace. More 96 THE NEOG^IC REALM. [CHAP. remarkable than any is the extinct Peltephilus of the Santa Cruz deposits, in which the teeth form a continuous series up to the front of the jaws, while the skull has a very broad snout, and the humerus is of such a remarkable shape that it has been described as that of a monotreme. Indeed this genus seems to suggest that edentates are derived from animals with a fully developed series of teeth in the front of the jaws. The pichiciagos (Chlamy- dophorus], which are unknown before the Plistocene, form a sub-family by themselves ; and yet another sub-family group is indicated by the gigantic Chlamydotherium, of the Brazilian caves and the Pampean, which rivalled the largest glyptodonts in size, and had teeth of a more complex type than the true armadillos. Other species occur in the Catamarca and Monte Hermoso Ter- tiaries, although the genus is unknown in those of Patagonia. From the armadillos and their immediate kin the extinct glyptodonts differ in having the carapace in the form of a con- tinuous solid shield, without any movable bands in the middle region ; in addition to which the skull is characterised by its depth and shortness, while the teeth form long fluted prisms. The internal skeleton, as shown in figure 19, is characterised by the union of nearly the whole of the vertebrae of the back into a solid girder for the support of the massive carapace ; and the feet are furnished with much shorter claws than those of the burrowing armadillos, the hinder ones being almost nail-like in form. Most of the species from the Pampean formation are animals of gigantic dimensions, the length of the carapace being not unfrequently from six to eight feet ; and they are unquestion- ably some of the most extraordinary creatures that ever trod the earth. Although the majority are South American, some members of the genus wandered as far north as Texas, while from the upper division of the Loup Fork beds, corresponding to the lowest Pliocene, a North American form has been described under the name of Carioderma. In the typical genus Glyptodon, which ranges from the sand- dunes and Pampean formation to the Monte Hermoso beds, the tail-sheath, as shown in the annexed figure, is composed of a number of spiny rings, gradually decreasing in size from the root to the tip, and the polygonal plates of the carapace are each I Cs. O 0 98 THE NEOG/EIC REALM. [CHAP. ornamented with a distinct rosette-like sculpture. The allied genus Plohophorus, of which the remains occur alike in the Brazilian caves and the Catamarca and Monte Hermoso beds of Argentina, while agreeing with the last in the characters of the skull, has a carapace more like that of the undermentioned Panochthus, and a tail-sheath resembling that of the next genus. In addition to well-marked distinctive features of the skull, Loma- phorus is characterised by the great elongation and slender form of the carapace, which is produced on either side of the neck in the same manner as in the armadillos, while its margins lack the large bosses exhibited by the typical genus. The tail-sheath consists of a small number of rings at the base, followed by a long terminal tube ornamented with smooth, oval, bony plates, of which those along the sides and at the tip are larger than the rest. The genus, of which the species are much inferior in point of size to those of Glyptodon, has the same geological range as Plohophorus. Another type of the family is represented by the animals constituting the genus Panochthus, in which the hexagonal bony plates of the carapace are arranged in more distinct rows on the sides of the body ; those of the back being ornamented either with a number of small granular tubercles (as in the species here figured), or with one circular central disc surrounded with several rows of much smaller discs. A more striking difference is dis- played in the structure of the sheath of the tail, which consists at the base of six or seven large smooth rings diminishing very rapidly in diameter, and terminates in a long and massive de- pressed tube, the sides of which are ornamented with large roughened bosses, probably surmounted during life with horny knobs, while the intervening spaces are covered with small bony ossicles. The species are of very large or medium size, and range from the Pampean to the Monte Hermoso beds in Argentina, while some occur in the Brazilian caves. Still more extraordinary is the gigantic Dadicurus of the Pampean, represented by a some- what smaller form from the Monte Hermoso beds. Having a total length in a straight line of close upon twelve feet, five of which are taken up by the ponderous tail, the Pampean repre- sentative of this genus has the outer surface of the plates of the III.] GLYPTODONTS. 99 7—2 100 THE NEOG^IC REALM. [CHAP. carapace smooth ; each being perforated by three or four large holes for the passage of blood-vessels, and the whole being pro- bably invested with a continuous leathery skin, instead of each disc bearing a separate horny shield. Commencing with a small series of enormous, narrow, hoop-like rings, the tail-sheath termi- nates in a long, massive, depressed, and nearly smooth club- shaped tube, at the extremity of which are a number of rough disc-like surfaces, apparently for the attachment of large horns. None of the preceding forms, which include all the largest members of the family, range below the horizon of the Monte Hermoso, Catamarca, and Parana beds, but the group is also represented in the Santa Cruz deposits of Patagonia, although the single species found there is of much smaller dimensions than any of its later cousins, the whole length of the carapace not exceeding about two feet. It is noteworthy that this earliest known glypto- dont, on which the unwieldy name of Propalczohoplophorus has been conferred, presents certain indications of affinity to the armadillos in the structure of its carapace, in which incipient movable bands may be detected on the margins of the middle region. In this small size of their earliest definitely known representative, the glyptodonts resemble the under-mentioned ground-sloths. Unless the aforesaid remains from the Oligocene Phosphorites of France should prove to belong to the group, we are at present totally in the dark as to whence both the glyptodonts and the armadillos originally came ; and it is, indeed, quite probable that, like the other members of the order, they may have origi- nated in South America (if not in an Antarctic continent) from some at present quite unknown mammalian type. How such creatures, which seem absolutely unassailable, came to be extermi- nated, is one of those questions which it appears quite impossible to answer. Although they have not hitherto been discovered in a fossil state, the sloths, constituting the family Bradypodidce, are just as characteristic of Neogaea as the two pre- ceding groups. Their habits, however, necessarily restricting them to the tropical forest-districts, their absence in a fossil state 1 must 1 A presumed fossil sloth was described from the Argentine, but the jaw on which it was founded proves to belong to the Megalotheriidce. III.] SLOTHS AND ANTEATERS. IOI not be taken as an indication that they did not exist during the Pampean epoch, since their remains are not likely to occur in the Argentine, although they might with more probability be looked for in Brazil. On the other hand, their specialised struc- ture makes it highly probable that they had not come into being at the date of deposition of the Santa Cruz beds. Of the dimen- sions of medium-sized monkeys, sloths are characterised by their short, rounded heads, and extremely long limbs, armed with very elongated curved claws ; in the genus Bradypus the latter being three in number on each foot, but in Cholcepus reduced to two in the fore feet. Their bodies are coated with very coarse ragged hair, and the tail is wanting. The teeth are oval prisms, some- what cupped in the middle of their grinding surfaces ; but in the last-named of the two genera the first pair in each jaw are larger than the rest, from which they are separated by an interval, and form tusks wearing against one another to oblique facets. Usually there are five upper, and four lower teeth on a side. The range of sloths extends from Mexico throughout the greater part of the forest-districts, although they do not appear to reach as far south as Paraguay. Likewise unknown in a fossil condition, the true anteaters, or Myrmecophagidce, constitute another exclusively Neo- .. ' . . i • , Anteaters. gseic family, with nearly the same geographical range as the sloths, but represented in Paraguay. So unlike are these ani- mals to sloths, that it is at first difficult to believe that there is any close relationship between the two, and it is largely due to the evi- dence of the ground-sloths referred to below that it has been possible to discover how close the connection really is. In place of being rounded and shortened, the skull in the present family is more or less elongated and slender, with the jaws entirely devoid of all vestiges of teeth, and the tongue long, cylindrical, and extensile. An equally striking difference obtains in regard to the structure of the limbs, the fore foot of the great anteater having five toes, of which the middle one is much more powerful than the others, while all except the fifth are furnished with strong claws. In walking, the outer side and part of the upper surface of the fore foot is applied to the ground ; but in the hind limbs the sole forms the support in the ordinary manner. Whereas sloths are 102 THE NEOG^IC REALM. [CHAP. highly specialised as regards the structure of their limbs, in the present group the greatest degree of specialisation shows itself in the skull, in the majority of the species. There are but three FlG. 22. TAMANDUA ANTEATER. members of the family, each representing a genus by itself, namely the great anteater (Myrmecophaga\ the tamandua (Tamandua], and the two-toed or little anteater ( Cycloturus] ; the latter being ex- clusively arboreal in its habits. The foregoing remarks on some of the structural features of the sloths and anteaters will the more easily enable si?th°sUnd" the reader to understand the peculiarities of the extinct group of ground-sloths. They have been divided into a large number of genera and several families; but the former may be considerably reduced, and the whole of them included in the single family Megalotheriidce. Ranging in the Argentine from the Santa Crucian to the Pampean epoch and the overlying sand-dunes, the family has a geographical distributional area including North America as far as Kentucky. The South American forms vastly exceed those of N. America in point of number; and whereas the latter are found only in deposits of upper Pliocene and Plistocene age, the former, as we have seen, extend downwards to the Miocene. The members of this family may be denned as edentates without a carapace, the skull and dentition of the general type of those of the sloths, and the III.] GROUND-SLOTHS. IO3 limb-bones and vertebrae like those of the anteaters. The skull is, however, somewhat more elongate than in the former, and in the case of the genus Scelidotherium approximates to that of the latter. The Plistocene forms include by far the largest representatives of the order, the Megalotherium1 attaining a total length of about eighteen feet, with a bodily bulk fully as great as that of an elephant. Whereas all the members of the family whose remains occur in the Plistocene walked on the outer sides of their feet, in the small ancestral Patagonian forms this specialised character seems to have been less developed. The typical genus Megalotherium — which includes several species, ranging in time from the Monte Hermoso and Cordoba beds to the Pampean, and in space from Argentina and Chili to South Carolina and Texas — is sufficiently distinguished by having the teeth in the form of large quadrangular prisms, sometimes measuring as much as a foot in length, and wearing on their sum- mits into a pair of transverse ridges, owing to the presence of layers of unequal hardness. The allied genus Mylodon, including smaller forms which may be compared in size to rhinoceroses, differs from the preceding in the structure of the teeth, which are similar to those of the sloths ; the skull, as shown in the figure on p. 104, being comparatively short, with the teeth extending nearly up to the extremities of the jaws. In the skeleton of this genus the limbs are of moderate length and very powerful. The two outer toes of the fore feet are rudimental and clawless, but the three innermost provided with claws, of which the third is much larger than either of the others, this discrepancy being carried to a still greater extent in Megalotherium. It will be observed that the creature walked on the outer sides and part of the upper surfaces of the fore feet after the manner of a sloth ; but, unlike the latter, only the outer sides of the hind feet were applied to the ground ; the great middle toe, which in Megalotherium carried a gigantic claw, not touching the ground at all. In the structure of their feet these animals are thus more like anteaters than sloths, although the hinder pair are of a somewhat more specialised structure than in the latter. It may be mentioned that the 1 The name Megatherium clearly requires amendment to Megalotherium. 104 THE NEOG^IC REALM. [CHAP. conformation of their teeth indicates that the ground-sloths were vegetable-feeders, and it is probable that they subsisted largely upon the young twigs and leaves of trees, which may have been FIG. 23. UNDER-SURFACE OF SKULL OF Mylodon. (Reduced.) brought within reach by the animals rearing themselves up against the trunks and pulling down the boughs with their fore paws. The present treeless condition of the Argentine pampas suggests that the ground-sloths were grazing rather than browsing III.] GROUND-SLOTHS. IO5 animals, but their structure is not in favour of this view, and it must be remembered that their remains are likewise met with in Brazil, which was probably always as well wooded as at the present day. The disappearance of forests from the pampas cannot, in- deed, be regarded as more marvellous than the extinction of its Plistocene mammals. In the sand-dunes near the coast at Buenos Aires bones of some of the ground-sloths, as well as of glyptodonts, have been found in association with human remains, so that their extinction is an event of comparatively recent date. The genus is typically represented by Mylodon harlani of the Plistocene of Kentucky and other parts of North America, but is nevertheless essentially South American, ranging in Argentina from the Pam- pean beds to those of Parana and Monte Hermoso. The allied genus Megalonyx is exclusively restricted to the North American Plistocene and Upper Pliocene ; and here may be repeated the observation that the absence of remains of these ground-sloths from the Miocene of North America, coupled with their presence in the Santa Cruz beds of Patagonia, clearly indicates that they are late immigrants from the south into the northern half of the continent. , Nearly allied to Mylodon, the genus Glossotherium from the Plistocene of Argentina and Uruguay serves to connect it with another generic representative of the family known as Scelido- therium. In place of the comparatively short skulls of the mylodons, the species of this genus have the muzzle of the skull greatly elongated, so that there is a long toothless space in advance of the dental series ; and whereas the skulls of the species of Mylodon are essentially sloth-like, those of Scelidotherium show a marked approximation to the anteater-type. The species of Scelidotherium are of medium or rather small size ; and in space the genus ranges from Patagonia, through the Argentine, to Brazil, Bolivia, and Chili ; while in time it extends from the superficial sand-dunes and Pampean deposits to the lower Tertiaries of Parana, Monte Hermoso, Catamarca, and Santa Cruz, with a gradual decrease in the size of the species as we descend in the geological scale '. Nearly allied is the genus Catonyx, from the 1 The Santa Cruz form has been quite unnecessarily separated under the name of Analcithermm. io6 THE NEOG^EIC REALM. [CHAP. ! w ^ K^* •-£ III.] MARSUPIALS. 107 Brazilian caverns ; while Nothrotherium of the same deposits seems to have been another nearly related form with teeth of the Megalotherium type. The imperfectly known Nothropus of the Pampean and Ortotherium of the Parana beds seem, on the other hand, to be late survivals of another group typically represented by the genera Eucholaops and Pseudhapalops of the Santa Crucian epoch of Patagonia. These forms differ from all those noticed above in that the terminal joints of some of the toes have a median cleft as in the great anteater, and likewise in the elongation of the metatarsal bones ; and it seems probable that the hind foot was not so much everted as in the later representatives of the family. The skull is of the general type of that of Mylodon ; most of the molars being prismatic in form, and surmounted by a pair of transverse ridges, more or less closely connected at their extremities so as to produce an oval cavity on the grinding surface. The first tooth is, however, tusk-like, and separated by a gap from the others. In some of these early ground-sloths the skull did not exceed three inches in length ; but other species were con- siderably larger. They are evidently generalised types, and were probably nearly allied to the ancestral stock which gave rise to Mylodon and Megalonyx, if indeed they be not the actual pro- genitors of both. The last group for consideration is that of the marsupials, or pouched mammals, among which the family of the opossums (Didelphyidce), with the three genera Didelphys1, Dromiciops, and Chironectes, is now confined to the New World, the great majority of the numerous species being Neogaeic, although the common opossum (Didclphys marsupialis} ranges from Chili and Brazil to the United States. Although certain forms from the Santa Cruz beds described under the names of Eodidelphys and Prodidclphys were originally assigned to the present family, these have been subsequently identified t^y Dr Ameghino2 with the under-mentioned family of the Microbio- theriidce. True opossums occur, however, in the Monte Hermoso 1 This genus is divided into several sub-genera, regarded by some writers as entitled to generic rank. 2 Bol. Ac. Cordoba, Vol. xm. p. 363 (1894). io8 THE NEOG^IC REALM. [CHAP. beds • while, as mentioned in the last chapter, they were widely distributed in the northern hemisphere during the Oligocene. If the conclusions of Dr Ameghino are right as to the absence of these marsupials from the Santa Cruz beds, it is evident that opossums only reached South America from the north at the close of the Miocene or commencement of the Pliocene, and that they do not belong to the indigenous fauna. It has been generally considered that the common opossum of the United States is a direct survivor from the Oligocene forms of North America, but it is more probable that it is really a very recent immigrant from the south, seeing that fossil representatives of the genus are unknown from the North American Miocene and Pliocene. During the Miocene the group perhaps survived in the extreme south of North America. Although opossums are apparently wanting, the Santa Cruz beds have yielded remains of undoubted marsupials, but several FIG. 25. LEFT HALF OF LOWER JAW OF Prothylctcinus, nat. size.) of them are assigned by Dr Ameghino to a distinct group under the name of Sparassodonta. Foremost of these is the genus ProthylacinuS) already mentioned in the last chapter, which may be provisionally assigned to the Australian Dasyuridce. In having only three in place of four lower pairs of incisors this genus agrees with the latter family, and differs from the opossums ; while the whole character of the lower jaw and dentition is very similar to that of the Tasmanian Thylatinus, with the exception that the premolars are closer together. As in the Dasyuridce generally, III.] MARSUPIALS. 109 there are four pairs of upper and three of lower incisor teeth in Prothylacinus, and the same is the case with the smaller Santa Crucian form described as Amphiproviverra, which appears to be of a distinctly dasyurid type, although not coming very near to any Australian genus. v With regard to the Microbiotheriida, as typified by the genus Microbiotherium, these, although they are not included by Dr Ameghino in the order, appear to be undoubted marsupials, since they have a dentition numerically the same as that of the opossums, vacuities in the palate, and an inflected angle to the lower jaw. From the opossums they differ by the non-production of the palate behind the last molars, and in the form of the lower jaw, in which the extremity is produced to a greater extent in advance of the canine. In all these respects they approximate to the Dasyurid genus Phascologale, from which they differ in having one pair less of incisors in each jaw. The ancestors of the Australian Dasyuridce. must, however, have originally had five pairs of upper and four of lower incisor teeth, as the former are retained in many of the ban- dicoots (Peramelida), while Myrmecobms occasionally develops four pairs of these teeth in the lower jaw. It seems therefore probable that the Microbiotheriidcz were minute polyprotodont marsupials of an Australian type. There is more difficulty in arriving at any satisfactory con- clusions as to the serial position of certain carnivorous mammals from the Santa Cruz beds, of which a large form described as Borhycena may be taken as an example. In these animals the dentition approximates to a certain extent to that of the primitive or creodont Carnivora of the earlier Tertiaries of the northern hemisphere, although retaining the marsupial feature of four pairs of molars and only three of premolars. The replacement of the teeth is also fuller than in the marsupials. Dr Ameghino has suggested that these animals were transitional between marsupial and eutherian carnivores, and that the latter group originated in South America ; but this idea is obviously untenable. A possible suggestion is that they may be specialised offshoots from the marsupial stock which died out without giving origin to any descendants. A small mouse-like mammal first described in 1863 upon the 110 THE NEOG^IC REALM. [CHAP. evidence of a single specimen from Ecuador under the name of Hyracodon fuliginosus, but whose affinities were not determined till 1895, when an example of a second and larger species was procured from Bogota, belongs to the sole surviving genus of a group of small marsupials which occur abundantly in the Santa Cruz beds, and were till quite recently regarded as extinct. Un- fortunately the name Hyracodon has been previously employed to designate an extinct ungulate, and it has accordingly been replaced by Ccenolestes. The essential characteristic of this group of marsupials, is that while their upper dentition is of a poly- protodont type, that of the lower jaw is very similar to the diprotodont modification. In the living species, for instance, there are four pairs of small upper incisors of a normal type, FlG. 26. FORE PART OF THE RIGHT HALF OF THE LOWER JAW OF Acdestis oweni. (Much enlarged.) The first tooth on the right is the first incisor, and the one on the extreme left the first molar. followed by a large canine, while in the lower jaw, as shown in the accompanying figure of that of one of the extinct forms, there is a large pair of horizontally projecting incisors, succeeded by several minute functionless teeth, of which the first three represent the second and third incisors and the canine. In all the forms the molar teeth have quadrangular crowns, surmounted by four blunt tubercles, somewhat resembling in structure those of certain ungulates, and thus totally different from the triangular and sharply-cusped molars of the opossums and other polyprotodonts. In the living forms, as well as in certain fossil kinds (Epanorthus, Decastis and Acdestis] from the Santa Cruz beds, the last premolar tooth, as shown in the figure of the jaw of Acdestis, is of normal dimensions : and these forms may consequently be grouped in a single family, under the name of Epanorthidce. In another group, III.] MARSUPIALS. 1 1 1 confined to the Santa Crucian horizon, where it is represented by the family Abderitidce, the last premolar in each jaw is much larger and taller than the other teeth, and has its crown in the form of a compressed cone, marked on the sides with vertical grooves, as exhibited in the figure of Abderites. A third family is FlG. 27. RIGHT HALF OF LOWER JAW OF Abderites, MUCH ENLARGED. known as the Garzoniidce. In all the skull is of an elongated form, with large vacuities both in the front and hinder part of the palate, and presents a considerable general resemblance to those of the Australian genera Peragale and Perameles. With the ex- ception of the retention of four pairs of upper incisors and the small size of all these teeth, the dentition exhibits, however, a remark- able approximation to that of the Australian diprotodont genus Dromicia. On the other hand, the feet are of normal structure, with five complete toes, none of which are united by integuments; the thumb and great toe being apparently slightly opposable to the other digits. Probably the rat-like tail is slightly prehensile at the extremity ; and a small pouch is present in the female. In the skeleton the lower jaw exhibits the usual inflection of the angle ; and the pelvis carries marsupial bones. Probably these Patagonian marsupials, which may be known as selvas, must be included in the diprotodont sub-order ; from the Australian representatives of which they differ by the small and numerous upper incisors and the non-syndactylous hind feet. Both these being generalised features, it is evident that if the selvas are true diprotodonts their ancestors must have originated from the polyprotodonts in Notogaea, for if they are of exclusively South American origin they must form a subordinal group by 112 THE NEOG^IC REALM. [CHAP. themselves. Assuming their affinities with the Australian type to be rightly determined, they constitute a most important link in the chain connecting the faunas of South America and Aus- tralia. In the last chapter it has been argued that, from the absence of allied forms in the Tertiaries of North America and Europe, as well as from their resemblance to the Australian dasyurids, it is difficult to come to any conclusion other than that the ancestors of the Santa Crucian polyprotodont marsupials reached the country from Australia, either by way of the Antarctic continent, or by a land-bridge in a more northern part of the Pacific. If this be correct, and likewise the supposition that the opossums origi- nated from the ancestral stock in South-eastern Asia, it will be evident that Didelphys and Ccenolestes met in South America after their ancestors had travelled half round the world in opposite hemispheres. It may be added that the alleged occurrence of monotremes in the Santa Cruz beds is due to bones of aberrant armadillos (Peltephilus) having been mistaken for those of that group1. Although in this volume the writer avoids laying much stress upon aquatic mammals, it may be mentioned that Cetaceans. there are two genera ot dolphins belonging to the family Platanistidce, each represented by a single species, which are peculiar to the Neogaeic realm. These are Stenodelphis (Pontoporia] from the mouth of the Rio de la Plata, and Inia of the upper Amazon; the only other existing representative of the family being Platanista of the larger Indian rivers. After the foregoing survey of the chief features of the recent and fossil mammalian fauna of the Neogaeic realm, tinction ofThe ^ts general bearings on the relations of South America to other parts of the world may be taken into con- sideration. It will, however, facilitate matters to give a tabular view of the orders, suborders, and families of non- volant land mammals represented in the realm. In the following table such groups as are either confined to Neogsea, or have only reached North America at a comparatively recent epoch are 1 See Lydekker, An. Mus. La Plata— Pal. Argent. Pt. III. p. 67 (1894). III.] LIST OF THE FAUNA. 113 printed in italics ; the extinct types being indicated by the prefix of an *, while those dating from the Santa Crucian (or the earlier Patagonian) epoch are followed by the words Santa Cruz, and those from the Parana beds by the word Parana. I. PRIMATES. — Santa Cruz. Cebida, — Santa Cruz (* Hdmunculus). Hapalidcz. II. CHIROPTERA. Phyllostomatida. — One genus ranging to Cali- fornia. Emballonuridae. — Seven peculiar genera. Vespertilionidae. — Natalus, Thyroptera. III. INSECTIVORA. Solenodontidce. — West Indies. IV. CARNIVORA. Felidae. — No peculiar genera. Canidae. — In addition to the cosmopolitan Canis, represented in Brazil by the peculiar Icticyon. Ursidae. — No peculiar genera; *Arctotherium, common to N. America. Procyonidae — Nasua, Cercoleptes, Bassaricyon, and * Cynonasua. Mustelidae. — Galictis, and Conepatus, the latter extending into Texas. V. UNGULATA. i. ARTIODACTYLA. Cervidae. — Cariacus, peculiar to New World. Dicotylidae. — Peculiar to New World at present day. Camelidae. — Lama. L. 8 114 THE NEOG^IC REALM. [CHAP. 2. PERISSODACTYLA. Tapiridae. — Elsewhere only in Malaysia at the present day, but formerly widely distributed over the northern hemisphere. Equidae. — Now unknown, except through intro- duction. In addition to the cosmopolitan Equus (including Tertiary forms), the pecu- liar Pampean genera * Hippidium and * Ono- hippidium. 3. * Litopterna.— Santa Cruz. * Macraucheniidce. — Santa Cruz. * Proter other iidtz. — Santa Cruz. 4. * Astrapotheria. — Santa Cruz. * A strapotheriidce. — Santa Cruz. * Homalodontotheriida. — Santa Cruz. 5. * Toxodontia. — Santa Cruz. * Toxodontidce. — Santa Cruz. * Typotheriida. — Santa Cruz. * Pachyruchidce. — Santa Cruz. 6. * Pyrotheria. — Patagonian beds lying below those of Santa Cruz. * Pyrotheriidte. — Patagonian beds. 7. PROBOSCIDEA. Elephantidse. — Represented by Mastodon in the Pampean and Monte Hermoso beds. VI. RODENTIA. I. SCIUROMORPHA. Sciuridae. — Represented by Sciurus as far south as Paraguay. * Castoroididse. — Peculiar to New World. 2 MYOMORPHA. Muridae. — Represented by species of the New World genus Sitomys, as well as by several peculiar types such as Rhithrodon, Ichthyomys of Peru, Holochilus of Brazil, etc. III.] LIST OF THE FAUNA. 115 3. HYSTRICOMORPHA. — Santa Cruz. Caviidce. — Paran a. Dasyproctidce . Dinomyidce. Lagostomatidcz. — Santa Cruz. Hystricidae. — Santa Cruz. Octodontidae. — Mainly Neotropical, but also Ethi- opian. Santa Cruz. * Eocardiidce. — Santa Cruz. 4. LAGOMORPHA. Leporidae. — Represented by two species of Lepus. VII. EDENTATA. — Santa Cruz. Dusypodida. — Santa Cruz. A few forms range as far as Texas. * Glyptodontida. — Santa Cruz. One Neogaeic genus ranges as far north as Texas, and a peculiar one has been described from further north. * Megalotheriidce. — Santa Cruz. Mainly Neogaeic, but also ranging into North America. Myrmecophagidcz. Bradypodidcz. VIII. MARSUPIALIA. — Santa Cruz. 1. DIPROTODONTIA. — Santa Cruz. (Elsewhere only in Notogaea). Epanorthidcz. Santa Cruz. (One existing genus, Ccenolestes). * Abderitidcz. — Santa Cruz. * GarzoniidcB. — Santa Cruz. 2. POLYPROTODONTIA. — Santa Cruz. Didelphyidae. — Now mainly Neogaeic, where they date from the Monte Hermoso stage, but ranging into North America, and formerly widely spread over the northern hemisphere. Chironectes, several subgenera of Didelphys, and Dromiciops peculiar to the realm. 8—2 Il6 THE NEOG^EIC REALM. [CHAP. VIII. MARSUPIALIA (continued}. Dasyuridae. — Santa Cruz. Now confined to Notogaea, but apparently represented in the Santa Cruz beds by * Prothylacinus and * Amphiproviverra. * Microbiotheriidcz. — Santa Cruz. Incertce, Seats. * Borhyanidce. — Santa Cruz. Although the addition of the names of all the peculiar genera, both recent and extinct, would have rendered the distinctness of the Neogaeic mammalian fauna still more pronounced, yet the foregoing table as it stands is amply sufficient to show that Neogaea is entitled to form one of the three primary zoological realms of the world. Starting with the Santa Crucian epoch of Patagonia and the somewhat older Patagonian stage, which form the earliest date at which the history of the Tertiary land mammals of Neogaea can be taken up, there is evidence that at least the southern part of the area was populated by the following pecu- liar fauna. Firstly, monkeys of a type quite different from those of the Old World, but evidently allied to the existing Neogaeic forms, were abundant; while rodents, belonging to the same groups as those now inhabiting the continent, several of which were nearly allied to existing African forms, and more remotely to certain Oligocene European types, attained a great development. Probably Insectivora with V-shaped molars were also present. More peculiar are the extinct subordinal groups of ungulates described above, which appear to have been allied to the ancestors of the perissodactyles of the northern hemisphere, and may possibly be remotely connected with the African hyraces. At the same period flourished several families of edentates (a group which in its restricted sense was originally peculiar to Neogaea), such as armadillos, glyptodonts, and ground-sloths, most of the members of which were of comparatively small size ; but of the ancestry of this group nothing can be said with certainty. Among the marsupials, although opossums appear to have been wanting, there were several types seemingly allied to Notogaeic forms, while others which may be included in the order were more III.] RELATIONS OF THE FAUNA. 117 or less unlike any from other regions. In addition to true mar- supials, the only carnivorous types were the problematical Borhycenidce. Now, as stated in the earlier part of the chapter (p. 68), this fauna cannot be older than the lowest Miocene or highest Oligocene ; and among its deficiencies may be noted lemuroids, true carnivores, creodonts, artiodactyle and perissodactyle ungulates, and opossums, all of which were in existence during the Oligo- cene or Miocene in North America and Europe. Moreover, at those epochs the former country lacked the whole of the Neogaeic types. Clearly, then, there must have been a barrier between North and South America during the Oligocene and a Early Separ- portion or the whole of the Miocene; but before ationofN.and entering into the consideration of other evidence S' Amenca- showing the nature of that barrier, it may be well to give a table of the mammaliferous Tertiary strata of North and South America, with their approximate European equivalents1. In descending order this runs as follows : Plistocene Up. Pliocene Low. Pliocene Miocene Up. Oligocene Mid. Oligocene Low. Oligocene Up. Eocene Mid. Eocene Low. Eocene Lowest Eocene South America. Pampean Santa Cruz Patagonian North America. Equus beds. Blanco Monte Hermoso f Palo Duro. (?) Parana Loup-Fork -j Nebraska tDeep River (Hiatus) John Day / Protoceras Beds White River j°reodon Beds Titanother- ium Beds Uinta rWashakii Bridger -I Bridger. IWind River. Wahsatch Puerco Europe. Cave-deposits etc. ? Crag. | Pikermi Sansan St Gerand-Le-Puy Ronzon. Montmartre Parisian Suessonian Cernaysian 1 In compiling this table the writer is indebted to Prof. W. B. Scott. Many American geologists (among them Dr Scott) include the whole of the Il8 THE NEOG^IC REALM. [CHAP. Regarding the geological evidence of a separation between the two Americas, Cretaceous marine strata occupy a large portion of Mexico; and in 1879 Dr Le Conte1 wrote as follows. The shore line of the Gulf of Mexico "was much more extended both north- ward and westward than either now or in Tertiary times. From the Gulf there extended northwestward an immensely wide sea, covering the Plains region and the Rocky Mountain region as far westward as the Wahsatch range, and dividing the continent into two continents, an eastern or Appalachian, and a western or Basin-region continent. Probably also this sea connected across the region of Mexico with the Pacific, thus dividing the western continent into two, a northern and southern." Later observations have shown that the Cretaceous sea undoubtedly made a wide gap between North America and the southern portion of the con- tinent2; while the existence of Oligocene or Miocene strata in the region of the isthmus of Parana shows that the separation, which probably continued through the Eocene, was in existence during Loup-Fork in the Miocene; while to the lower part of the same era they assign the John Day beds of America, and the St Gerand-le-Puy beds of Europe. Others (e.g. Prof. Osborn, Studies Biol. Labor. Columbia Coll. vol. i. pt. 2, p. 28, 1893) refer the Equus beds of North America to the Pliocene. The following quotation from a paper by Prof. Cope (American Naturalist, 1895, p. 599) conclusively proves their Plistocene age. There it is stated that "the Equus beds are found covering areas of various extent in Oregon, Nevada, California, the Staked Plains, Southern Texas, Chihuahua and the valley of Mexico. Their most eastern station is western Nebraska. They contain a fauna which includes one extinct species (Equus major Dek.) of the Megalonyx fauna, and the recent Castor fiber. They contain the extinct genus of sloths Mylodon, of a species different from that of the east, and four species of camels of the extinct genus ffolomeniscns, and a peccary. Recent species of Cam's and Thomomys occur, while two extinct horses (Equus occi- dentalis Leidy and E. tau Owen) are common. The hairy elephant (E. primi- genius) is abundant, while the Mastodon americanus is rare, if occurring at all. The proportion of recent to extinct species and genera in the Equus bed fauna is very similar to that occurring in the Megalonyx fauna, while they differ as to details." 1 Elements of Geology, pp. 451, 452, New York (1879). 2 This separation also existed in the Jurassic era, when, as shown by Neumayr (Erdgeschichte, 2nd ed. vol. II., p. 263), South America was united across the Atlantic area with Africa and Madagascar. I III.] RELATIONS OF THE FAUNA. 119 the middle portion of the Tertiary epoch1. When the connection between North and South America was completed is not precisely fixed by geological evidence ; but the occurrence of a glyptodont in the Nebraska stage of the Loup-Fork group, shows that it must have been by the end of the Miocene2. The question of a con- nection between the two continents by way of the West Indies is discussed later in this chapter, where it is concluded that if such a connection existed at all, it must have been of a transient nature. Having thus shown, both from palaeontological and geological evidence, that the early mammalian fauna of Incursion of Neogaea appears to have been totally isolated from Northern that of North America up to about the end of the Miocene, the question of the origin of that fauna may be deferred, and the irruption of northern types after the connection between North and South America had been established taken into con- sideration. It may be mentioned, however, that it was not till after this irruption of the northern forms that the essentially Neogaeic fauna attained its maximum development in respect to the bodily size of its constituents ; since a gradual increase in this respect maybe traced from the small glyptodonts and sloths of the Santa Cruz epoch, through the larger ones of the Monte Hermoso horizon, to the gigantic forms characteristic of the Pampean and the cavern deposits of Brazil. The presence of a glyptodont in the Nebraska stage of the Loup-Fork group in North America, and of northern forms in the Monte Hermoso horizon of South America, marks, then, the first commingling of the original faunas of the two halves of the New World. For the first time in the history of the southern continent this connection allowed of the immigration from the north of the true Carnivora, such as the existing cats (Felts) the extinct sabre- toothed tigers (Macharodus), dogs and foxes (Canida), bears (Ursus and Arctotherium\ raccoons (Procyonidtz), skunks and their allies (Mustetida), together with various ungulates belonging to sub-orders previously unknown in the realm. These latter include the guanaco and vicuna (Lama] — of which ancestral forms 1 See J. W. Gregory, Quart. Joum. GeoL Soc. Vol. LI. pp. 299, 300 (1895). 2 As stated above, many refer the whole Loup-Fork group to the Miocene. I2O THE NEOG^EIC REALM. [CHAP. are abundant in the North American Tertiaries — New World deer (Cariacus], horses (Equidce) of various genera, tapirs (Tapirid&\ peccaries (Dicotylidte), and mastodons. Among the rodents, squirrels, the various genera of Muridce, and the hares, likewise at this epoch made their first, appearance on the scene. Opossums also at this time effected an entrance into the land which has now become their chief home. That this new fauna came in from North America, and not from any other part of the world, may be regarded as certain from the presence of such essentially New World types as raccoons and their allies, skunks, peccaries, Cariacus, and Camelidce (exclusive of the Old World genus Came/us, which is of late origin), coupled with the absence of true deer (Cervus), pigs (Sus), Old World monkeys, and lemurs. At the same time this union of the northern and southern halves of the New World allowed certain members of the original Neogaeic fauna to make their escape into North America, glypto- donts, as already said, making their appearance in the Nebraska stage of the Loup-Fork group of the United States, while the ground-sloth Megalonyx occurs in the Blanco Beds. Although it is not universally admitted1, there is some evidence to indicate that this land connection was of com- paratively brief duration, seeing that none of the characteristic extinct types of South American ungulates, nor any of the peculiar Neogaeic rodents, reached the northern half of the continent. During the whole time that the alluvial deposits of the Parana and Paraguay rivers were being laid down, and well on into the human period, the mammalian fauna of the Pampean epoch, formed by an admixture of southern and northern types, continued to flourish, until the time when there came a complete sweep of all the larger forms, clearing off the whole of the ground-sloths, glyptodonts, mastodons, toxodonts, macrauchenias, horses, sabre- toothed tigers, and the larger members of the camel tribe, and in the Argentine leaving only armadillos, guanacos, a few deer, a number of rodents, various cats and foxes, as well as skunks and certain other members of the weasel family, to represent the vast assemblage of strange and giant creatures that once roamed over 1 See Gregory, op. cit. p. 300. III.] DISAPPEARANCE OF LARGE FORMS. 121 its plains. With regard to this extraordinary, and apparently sudden disappearance of almost all the larger forms of animal life from South America, it may be pretty confidently asserted, from the organisation of the creatures themselves, that at the time when the ground-sloths flourished, extensive portions of what is now the open pampas of Argentina were covered with forest ; and why the whole district should have become practically treeless, seeing that trees like the Australian eucalypti grow, when introduced, with more vigour than in their native home, is exceedingly hard to understand. That the country even when thus denuded was unsuited to the needs of the larger forms of mammalian life, is, however, negatived by the circumstance that in many parts the horses and cattle introduced from Europe have run wild and increased to an almost unprecedented extent. Neither does it appear that the extermination can be attributed to a period of extreme cold, since a glaciation of the pampas would assuredly have left unmistakable evidence of its presence. It is likewise practically certain that the clean sweep of the forests of Argentina and the larger mammals of the whole of South America is not due to the hand of man. It has, indeed, been suggested that the vast herds of guanaco which formerly roamed the pampas may have cleared the forests by preventing the growth of the seedlings ; but when we recall the fact that numbers of this group of animals flourished during the period when the alluvial formation was in the course of being deposited, it scarcely looks as though this could have been a vera causa. Moreover, if the forests were by some means or other actually destroyed, and the extermination of their animal denizens thus encompassed, there would still remain the disappearance of plain-dwelling forms, like horses, to be accounted for. Some have thought that pumas, by preying on the colts, were the active agents in this instance ; but even if such were really the case, it gives no help with regard to the disappearance of the ground-sloths and glyptodonts; — the latter being such strongly armoured creatures that it is absolutely certain they were not killed off by any animal foes. The problem is further com- plicated by the circumstance that the fossil remains of nearly all the larger animals which formerly inhabited the pampas are also found in the caverns of Brazil, where the climate is now, and 122 THE NEOG^IC REALM. [CHAP. probably always has been, tropical. Up to the present, it is, accordingly, impossible to account satisfactorily for the disappear- ance of all the larger forms from among the mammalian fauna of South America. Returning to the fauna itself, it may be asserted that before the great intrusion of northern forms the mammals of Neogaea were far more distinct from those of the rest of the world than is the case with the fauna of any other region, with the exception of Australasia; and that consequently there can be no hesita- tion in allowing this part of the earth's surface to take rank as a primary realm. At the time when the Santa Cruz fauna was so decidedly marked off there was a much more general similarity between the faunas of all the other regions of the world (exclusive of Notogaea) than is the case at the present day, and it is this antiquity of the differentiation of the Neogaeic fauna that supports so strongly its claim to distinctness. It has been suggested that the first land-connection between South and North America was probably of limited extent or short duration ; and some evidence of a later separation between the two areas is afforded by the beds of marine shells already mentioned as occurring in the upper part of the Pampean deposits ; these beds marking an epoch of submergence of a considerable portion of the area1. Subsequently to this final submergence of portions of Argentina and Uruguay there was a great upheaval of the country, indicated not only by the upraising of the aforesaid marine beds, but likewise by that of the sand-dunes fringing most parts of the Argentine coast. This upheaval, which has taken place within the human period, certainly resulted in the final union of the two Americas ; and since its date there has probably continued to be a greater and greater admixture of the two originally distinct faunas, so far as climatic conditions have permitted. It is, how- ever, not a little curious that some of the original northern types, such as the vicunas and guanacos, have entirely died out in their original habitat, to flourish only in the southern half of the con- tinent. 1 There is some evidence to show that the isthmus of Panama was never completely submerged after the Pliocene, but it may have been so narrow as not to allow of much migration of the fauna. III.] DISTINCTNESS OF EXISTING FAUNA. 123 Hitherto especial stress has been laid on the fossil mammals of Neogaea as entitling the tract to form a primary Distinctness realm, on account of the distinctness of its fauna of the existing from that of the rest of the world during a consider- able portion of the Tertiary epoch. Even at the present day, however, when, as already shown, its mammalian fauna contains a very large admixture of types which have immigrated from the north at a comparatively recent epoch, it still stands widely apart from other regions. On this point may be quoted the admirable summary given in "Island Life"1 by Dr Wallace, who writes that among the peculiar mammals we have " the prehensile-tailed monkeys and the marmosets, the blood-sucking bats, the coati- mundis, the peccaries, the llamas and alpacas [vicunas and guanacos], chinchillas, the agutis, the sloths, the armadillos, and the ant-eaters ; a series of types more varied and more distinct from those of the rest of the world than any other continent can boast of. Among birds we have the charming sugar-birds, forming the family C&rebidce, the immense and wonderfully varied group of tanagers (Tanagridce), the exquisite little manakins and the gorgeously-coloured chatterers ( Cotingidce) ; the host of tree- creepers of the family Dendrocolaptidce, the wonderful toucans (Rhamphastid Eastern which clearly show that the communication between the two areas was always more or less limited. In this chapter attention will be first directed to some of the most striking peculiarities of the mammalian fauna of Eastern Arctogaea, after which the whole fossil fauna may be taken into consideration. In addition to the total absence of existing opossums (Didel- phyid(E\ and the presence in its warmer portions of fruit-bats (Pteropodidce), which, however, are common to Notogaea, Eastern Arctogaea is especially characterised as being the home of all the higher Primates; namely the family Stmiidce, which includes the man-like apes and gibbons, and the Cercopithecidcz, embracing all the other Old World monkeys. From the South American monkeys (Cebidcz) both these families are broadly distinguished by having two pairs of premolar and three of molar teeth, whereas in the former group there are three pairs of both premolar and molar teeth. Not only are the two families in question confined at the present day to the Eastern hemisphere, but the same appears to have been the case at all epochs, since no trace of a fossil monkey has ever been recorded from North America. This remarkable isolation of the distributional areas of the Simiidcz and Cercopi- thecidce on the one hand, and of the Cebidce. (and Hapalidcz) on 12 — 2 ISO EASTERN ARCTOG^A. [CHAP. the other, points unmistakeably, in spite of their external similarity, to the dual origin of the monkeys of the Old and New Worlds. Both may, however, have originated from different groups of lemuroids, which, as indicated in an earlier chapter, ranged over the whole of Northern Arctogaea during the earlier part of the Tertiary epoch. At the present day the man -like apes, which are few in number, have an extremely limited distribution. The chim- panzees (Anthropopithecus) are restricted to Equatorial Africa, the gorilla (Gorilla) is found only in the hottest regions of Western Africa, the orangs (Simla} are confined to the islands of Borneo and Sumatra, and the smaller gibbons (Hylobates) are inhabitants of South-eastern Asia, from Assam and Burma to Hainan. Ex- tinct species of chimpanzees and orangs occur, however, in the Pliocene of Northern India; while gibbons are met with in the Miocene of France and Baden, although there is some difference of opinion whether these are generically identical with the Asiatic forms, or should be assigned to a genus apart (Pliopithecus}. The former deposits have also yielded remains of a large extinct ape (Dryopithecus)) apparently of a somewhat more generalised type than all the existing forms. The ordinary monkeys and apes (Cercopithecidce) have a wider distribution, ranging over most of the warmer parts of Eastern Arctogaea, and being represented by a single species at Gibraltar, and by two others in Moupin, in Eastern Tibet, and a fourth in Japan. This family, by the way, is not exclusively Arctogaeic, since, as we have seen, one species of a peculiar genus (Cynopi- thecus) inhabits Celebes. Apart from the latter, the family is represented by eight living genera, among which the Ethiopian Papio has extinct representatives in the Pliocene and Plistocene of India, as have also the Asiatic Macacus and Semnopithecus. The two latter genera also occur in the Pliocene of France and Italy; and a tooth of a species of Macacus has been obtained from the Plistocene brick-earths of Essex. In addition to these, there are certain extinct genera from the European Tertiaries; among which Mesopithecus from the lower Pliocene of Greece agrees with Macacus in its short and stout limbs, but approximates to Semnopithecus in the character of its skull and dentition. Dolicho- V.] DISTINCTIVE MAMMALS. l8l pithecus, from the French Pliocene, has a longer muzzle; while Oreopithecus, from the Italian Miocene, seems to connect the CercopitheddcB with the Simiida. It is thus evident that during the latter portion of the Tertiary epoch monkeys and apes were spread over the greater part of Eastern Arctogaea; and their extensive diffusion is a proof that this half of the realm could only have been connected with North America by land situated so far north that it formed an impass- able barrier to these animals. Although the smaller extinct European monkeys do not necessarily indicate a very high temperature in the regions they inhabited, there can be little doubt that at the era when Dryopithecus flourished southern Europe at least enjoyed a moist tropical climate. Not less characteristic of Eastern Arctogaea are the existing lemuroids, of which there are three families ; the largest (Lemurida) ranging over the Oriental, Ethiopian, and Malagasy regions, the Tarsiidce, with a single genus, being exclusively Oriental, while the sole representative of the Ckiromyida is Malagasy. The numerous existing members of the Lemuridce. are all characterised by the first of the three lower premolar teeth assuming the form and functions of a canine; and as this feature is unknown in any of the Tertiary representatives of the sub-order, this family appears to be exclusively confined to the area under consideration. For the most part, the Oligocene lemuroids of Europe seem likewise to have been markedly distinct from those of North America. Microchcerus^ , for instance, which is represented both in France and England, indicates a family characterised, among other features, by the general presence of only three pairs of premolar teeth in each jaw ; and Adapts, which is likewise common to the same two countries, has four such teeth. Although there are several families of Insectivora peculiar to the eastern hemisphere, the only one of these with a wide distri- bution is that of the hedgehogs and their allies (Erinaceida), which has representatives in the eastern Holarctic, Oriental, and Ethiopian regions. Although none are known from America, extinct repre- sentatives of this family are common in the European Oligocene. 1 Teeth figured on p. 156. 1 82 EASTERN ARCTOG^EA. [CHAP. Of these Pal&oerinaceus appears to be allied to the true hedgehogs (Erinaceus) ; whereas other genera, such as Necrogymnurus, connect the former with the existing long-tailed and spineless Gymnura of the Malayan islands. This group is thus essentially characteristic of Eastern Arctogaea as a whole. Turning to the Carnivora, we have, in addition to the Prote- leidce, of which the sole representative is the African aard-wolf, two important families practically confined to this half of the realm. The first of these is the extensive group of the civets, mungooses, and their allies ( Vtverridce), which has no representatives in the New World, although a single species in two genera ranges into the Austro-Malayan region. Out of a total of twenty-three genera included in this family only one mungoose (Herpestes) and the common genet (Genetta) range into Europe, most of the other forms being confined to the Oriental, Ethiopian, and Malagasy regions. Civets ( Viverra) and ichneumons, together with several remarkable extinct genera, such as Stenoplesictis, were, however, common in the European Tertiaries, from the lower Oligocene upwards. And from the circumstance that the last-named genus presents characters connecting the Viverridce with the weasel tribe (Mustelida), it would seem probable that the latter family was originally evolved in Eastern Arctogaea, although it has now a considerable number of American representatives. By means of certain extinct forms from the lower Pliocene of Southern Europe and Northern India known as Ictitherium, the civets are very closely connected indeed with the hyaenas (ffy&nidce) ; the three living representatives of which may be included in the single genus Hyana. Although the striped hyaena (H. striata) is now confined to Southern Asia and Northern Africa, it occurred in the Pliocene epoch in France and England; while the larger spotted hyaena (ff. crocutd] of Southern Africa ranged during the Plistocene era over the greater part of temperate Europe, and likewise extended eastwards as far as India. Numerous extinct species of the same genus are found in the Pliocene of Europe and India; and two extinct genera from the same deposits — Hyanictis and Palhyizna — connect the living forms with the aforesaid Ictitherium. Although one extinct Tertiary North American genus has been tentatively assigned to V.] RODENTS. 183 it, the family is thus essentially an Eastern Arctogaeic one; and it may be assumed that, as its living representatives are inhabitants of hot climates, the extinct forms were unable to exist sufficiently far north to permit them to cross by the bridge of land via Bering Strait. Among the rodents, in addition to the two widely spread families — Myoxidcz and Spaladdce. — confined to Eastern Arctogaea, we find in the Muridce the sub-family of the gerbils (Gerbillina) similarly restricted, their range including the whole of Eastern Arctogaea with the exception of the Malagasy region. The subfamily Murtna, which includes the true rats and mice (Afus) is likewise restricted to the Old World. These rodents differ from the hamsters (Cricetus) and the New World white- footed mice (Sitomys), which, with other forms, constitute the sub-family Cricetince, by the molar teeth in the upper jaw having their tubercles arranged in three longitudinal rows ; whereas in the latter they form only a double series. Distributed over all the regions of Eastern Arctogaea, with the exception of Madagascar, this group is also represented in the Australian region. Of the two extensively distributed families restricted to the area under consideration, the first is that of the dormice (Myoxidce) whose range includes the Eastern Holarctic and Ethiopian regions. Distin- guished from all other rodents by the absence of a caecum or blind appendage to the intestine, and further characterised by the complicated infoldings of the enamel on the crowns of their molar teeth, these beautiful little creatures now attain their maximum development in Africa. In a fossil state they are first known from the lower Oligocene of Europe, and are likewise common in the Miocene. Another family which does not range beyond the limits of Eastern Arctogaea is that of the Spalacida, typically represented by the great mole-rat (Spalax typhlus] — a blind creature, ranging over south-eastern Europe, Persia, Meso- potamia, Syria, and Egypt. The allied genus (Rhizomys), in which the eyes, although minute, are not covered with skin, includes several species, whose distributional area embraces the north of India, Tibet, China, Burma, Malaysia, and Abyssinia ; but the three remaining genera are restricted to the Ethiopian region. The whole of the foregoing belong to the mouse-like, or Myo- 1 84 EASTERN ARCTOGvEA. [CHAP. morphous section of the order, but among the Hystricomorphous rodents we have the typical porcupines (Hystricince), in which the tail is never prehensile, practically confined to Eastern Arctogsea, where they range over south-eastern Europe, and the Ethiopian and Oriental regions. The Javan species of the typical genus (Hystrix javanicd] is, however, found in the island of Timor, in the Austro-Malayan region, although it is doubtless of late introduction there, and may not improbably have been transported by human agency. In a fossil state porcupines of this sub-family are common in the European Tertiary as far down as the lower Oligocene. Turning to the ungulates, we have in the artiodactyle section two closely allied families, which — if we except certain pigs from the Austro-Malayan region and Papua, which may have been originally introduced by man — are restricted to the area under considera- tion. Both these families, moreover, have representatives, either living or extinct, in all the regions of Eastern Arctogsea, inclusive even of Madagascar, so that they may be reckoned among its most characteristic mammals. The Hippopotamidtz — now restricted to the Ethiopian region, where they are represented by the widely- spread common Hippopotamus amphibius, and the much smaller terrestrial H. liberiensis of the West Coast — ranged during the Plistocene and upper Pliocene epochs over the greater part of Europe as far north as England ; one species from these deposits being apparently indistinguishable from the common African form. Extinct species are met with in the Plistocene of Algeria, the Plistocene and Pliocene of India, the Pliocene of Burma, and the superficial deposits of Madagascar; some of these differing from the common hippopotamus by the presence of three, in place of two, pairs of incisor teeth in each jaw. Whatever may have been the case with the swine, it is evident that the hippopotami were never able to exist sufficiently far north to cross by way of Bering Strait into the New World. In the pigs (Suidez) — which among other features differ from the Hippopotamidce by the nostrils being perforated in a fleshy disc at the extremity of the muzzle, and like- wise by the structure of the teeth — the typical group of the genus Sus ranges over most of the Eastern Holarctic and the whole of the Oriental region, being replaced in the Ethiopian and V.] UNGULATES. 185 Malagasy regions by the potamochcerine group, frequently reckoned as a distinct genus. The Ethiopian region is, however, the sole habitat of the wart-hogs (Phacochoerus}. Among extinct artiodactyles we have two well-marked families, distinguished from the foregoing by the crescentic columns of their short-crowned cheek teeth — the Anoplotheriidce and Dicho- dontidce — likewise confined to Eastern Arctogaea, although their remains have hitherto been obtained only from the eastern section of the Holarctic region. The first of these includes several genera from Oligocene strata, characterised by having three columns on the front half, and two on the hinder half of their upper molars. In the typical Anoplotherium there were forty-four teeth, arranged FlG. 44. LAST FIVE RIGHT UPPER CHEEK-TEETH OF AN Anoplotherium. in a continuous even series, and the feet were provided with either three or two toes. Dacrytherium differs by the molars being more like those of Ancodus, and also by the deep cavity for the reception of a gland on each side of the face in front of the eye ; while the small and elegantly formed animals described as Xiph- odon have the crowns of the first three premolar teeth elongated and trenchant, the feet being two-toed. In the Dichodontida, of which there are likewise several genera, the cheek-teeth are more completely selenodont, with only four columns on the crowns of the upper molars ; and it is not improbable that in this family we have the ancestral types of both the chevrotains and the deer. In the Camelidce, as we have already seen, the typical genus Camelus, which is found living (although not in a wild state) in the Eastern Holarctic, Oriental, and Ethiopian regions, and fossil in the Plistocene of Algeria and the Pliocene of India, is likewise confined to Eastern Arctogsea. And the same is true, both in the 1 86 EASTERN ARCTOG^A. [CHAP. recent and fossil state, with the genera Tragulus and Dorcatherium, which at the present day alone represent the Tragnlida. The family of giraffes (Giraffidce), of which the Ethiopian Giraffa cameiopardalis is the sole existing survivor, was formerly extensively distributed over the area under consideration, to which it appears to have been always restricted, albeit represented by a considerable number of generic types. True giraffes (Giraffd) ranged during the Pliocene epoch over Greece, Persia, India, and China, and allied types are to be found in Visnutherium of the Pliocene of India and Burma, and Helladotherium from the corresponding formation of Greece. Still more gigantic than the latter were the huge Hydaspitherium, Bramatherium, and Sivatherium, of the Indian Pliocene, in all of which the simple horns of the giraffes were replaced by large antler-like appendages, differing considerably in their arrangement in the different genera. Other members of the family are Samotherium, of the Pliocene of the Isle of Samos, and Palceotragus from the equivalent deposits of Attica, in both of which the females appear to have been horn- less, although the males had a pair of simple, compressed, and nearly upright horn-cores. The former is represented by a species rivalling the giraffe in the size of its skull, but the latter was a much smaller animal. This group likewise extended to Northern Africa, where a large species from the Algerian Pliocene has been described under the name of Libytherium. Although the extensive family of the Bovidce, including the oxen, sheep, goats, antelopes, etc., is now represented in the northern part of the western Holarctic region by the American bison (Bos americanus\ the bighorn (Ovis canadensis], the musk- ox (Ovibos moschatus], and the so-called Rocky Mountain goat (Haploceros montanus), together with a few extinct forms from the superficial deposits, while the anoa (Bos depressicornis) is peculiar to Celebes, the greater number of its representatives belong to Eastern Arctogsea. The whole of the numerous genera of ante- lopes, together with the true goats, as well as the great majority of the sheep, are, for instance, restricted to this area. Moreover, whereas the musk-ox is now solely North American, it was common in Europe during the Plistocene ; while the bighorn is closely allied to the Kamschatkan wild sheep (O. nivicola), and the V.] EFFODIENTIA. l8/ American bison not far removed from its Caucasian cousin (Bos bison], so that all these forms are probably descended from ancestors inhabiting Eastern Arctogaea. In the perissodactyle section of the ungulates, if we take fossil forms into account, there are no families peculiar to this area; but among extinct forms we have the large Oligocene genus Palceo- therium1, and the Eocene Lophiodon absolutely restricted to it. Although occurring only in Syria and the Ethiopian region, and at present unknown in a fossil state, the peculiar subordinal group of ungulates represented solely by the hyraces (Procaviidce) per- haps deserves mention among the types characteristic of Eastern Arctogaea. A nearly similar observation applies to the extinct proboscidean family Dinotheriidce, in which the single known genus (Dinotherium) ranges from the Miocene and Pliocene of Europe to the Pliocene of Northern India. Of the edentates, with the exception of certain very doubtful forms from the French phosphorites, which may prove to be reptilian, we have no evidence of the existence of any representa- tives in the Old World. There are, however, in Eastern Arcto- g?ea two very peculiar families commonly assigned to the same order as the latter, although it seems preferable to regard them as indicating an ordinal group (Effodientia) by themselves. Of these the pangolins (Manidce), which are distinguished from all other mammals by their covering of overlapping horny scales, are now confined to the Oriental and Ethiopian regions, to which the one living genus Manis is common ; but they appear to have been represented in the Oligocene phosphorites of France by smaller extinct forms, to which the names Necromanis and Leptomanis* have been given. The second family, Orycteropodidce, of which the only living members are the Ethiopian aard-varks (Orycteropus), differ very widely from the last, the body being nearly naked, and the molar teeth characterised by a peculiarly complex structure which is unique in the whole mammalian class. A fossil species of the existing genus has been discovered in the lower Pliocene of Samos and Maraga, in Persia ; while the extinct genus Pal&orycte- 1 Teeth figured on p. 166. 2 The so-called Palceomanis, from the Pliocene of Samos, turns out to have been founded on remains of an ungulate. 1 88 EASTERN ARCTOG^EA. [CHAP. ropus, from the French phosphorites, is believed to have belonged to the same family, so that both groups appear formerly to have been widely distributed. Summarising the results of the foregoing survey, we may put in a tabular form the leading features of the mammalian fauna of Eastern as distinct from that of Western Arctogsea. In the sub- joined table the letters E., M., O., H. respectively indicate the Ethiopian, Malagasy, Oriental, and Eastern Holarctic regions ; and when a family is represented in any of such regions only in a fossil state, a f is added to the denoting letter. The names of such families or groups as are practically peculiar to the area 'under consideration are printed in italic type ; while extinct groups have an * prefixed. PRIMATES. Simiidce. O. E. H.f Cercopithecidcz. O. E. H. ; also extending into the Austro- Malayan region. Lemuridtz. O. E. M. * Microchoeridce. H. * Adapida. H. CHIROPTERA. Pteropodidse. O. E. ; also common to Notogaea. INSECTIVORA. Erinaceidce. O. E. H. CARNIVORA. Viverridce. O. E. H. M.; two species extending their range into the Austro- Malayan region. Hycenidcz. O. E. H.f RODENTIA. Muridae ; the sub-family Gerbillince, O. E. H., is restricted to Eastern Arctogaea, while the Murinae are exclu- sively confined to the Old World, but range into Notogaea. Myoxidce. H. E. Spalacida. H. O. E. V.] SUMMARY OF THE FAUNA. 189 Hystricidae ; in this family the Hystricina are practically restricted to Eastern Arctogaea, although a Javan species is found in Timor. UNGULATA. Hippopotamidce. E. O.f H.f M.f Suidce. H. O. E. M., also extending into the Austro- Malayan region. * Anoplotheriida. H. * Dichodontidce. H. Camelidae ; in this group the genus Camelus, H. E. O., is peculiar to Eastern Arctogaea. Tragulidae ; the existing genera Tragulus, O., and Dorca- therium, E. H.t, as well as several extinct ones, are restricted to this area. Giraffida. E. O.f H.f Bovidae; the whole of the true antelopes and goats, as well as most of the sheep and oxen, are restricted to Eastern Arctogaea, North America now preserving only one species of Bos, Ovis, Ovibos, and Haplo- ceros. * Palaeotheriidae ; Palceotherium. * Lophiodontidae ; Lophiodon. Procaviida. E. H. * Dinotheriida. H. O. EFFODIENTIA. Manida. E. O. H.f Orycteropodida. E. H.f To these may be added the absence of living opossums (Didel- phyida}. It will be observed that in this list such existing families as are confined to one of the regions of the area considered are for the most part omitted. Tertiary families which are at present unknown beyond the Eastern Holarctic region have, however, been included, since the limitation is probably in great part due to our want of knowledge of the early Tertiary faunas of the other regions. Although the differences indicated between the faunas of 190 EASTERN ARCTOG^A. [CHAP. Eastern and Western Arctogaea, which will be still more apparent after the consideration of the mammals of North America, seem at first sight to indicate that these two areas should form distinct divisions of the realm, yet the community of the fauna of the northern portion of the two hemispheres forbids this view. This question may, however, be more fully discussed in the chapter devoted to the Holarctic region. Before entering upon the consideration of the different zoolo- gical regions into which the realm is divided, it is MlmmarHan essential to take a brief survey of the Tertiary mam- Faunas of malian faunas of Eastern Arctogaea. By this alone Eastern . Arctogeea. it is possible to understand the true relations of the existing faunas to one another ; while such a survey also serves to demonstrate that the regions in question are but features of the present epoch of the earth's history ; and that even as late as the Pliocene portion of the Tertiary epoch the dis- tinctions now obtaining between the Holarctic, Oriental, and Ethiopian regions had no existence. In our survey we may omit the Eocene period, and commence with the lower Oligocene ; and it will simplify matters to give lists of some of the more important and better known generic types characterising the faunas of the different horizons. The leading affinities of many of the genera mentioned have been already alluded to in the present or preced- ing chapters ; but it would vastly exceed the limits of our space to attempt to point out the distinctive features of the others. Accordingly, the reader must either take them on trust, and treat them practically as abstract terms, or he must refer to some palaeontological treatise in order to find the real nature of the animals indicated by such generic names. Although it is essential to our purpose to notice the Oligocene faunas of Eastern Arctogaea, it is important to FauIS*06116 observe that our knowledge of these is practically limited to Western Europe. We are consequently quite unable to say how far the geographical range of such faunas extended, although it is probable that this embraced a large portion of the Eastern Holarctic region. Whether, however, at this epoch Ethiopian Africa had received a large mammalian fauna must be left for future discoveries to determine. V.] OLIGOCENE FAUNA. Ipl Under the term of lower Oligocene (the upper Eocene of many writers) are included a large series of strata, such as the freshwater beds of Bembridge and Hordwell in the south of England, the gypsum of Montmartre near Paris, and the corre- sponding black lignite beds of De'bruge in Vaucluse1. A consider- able part of the fauna of the Quercy phosphorites of Central France likewise comes under the same category, only we have here a mixture of Middle and Upper Oligocene forms. And in the case of the siderolites, or bone-earths of Switzerland, this admixture is carried to a still greater degree, undoubted Eocene types occurring with those properly characteristic of the Oligocene. In the following list such genera as are found only in the phosphorites have the letter P. after them; while after those peculiar to the siderolites the letter S. is added; both letters being given when the genera are common to the two formations. As already said, only some of the better-known forms are selected. Among the lemuroid Primates, we have the genera Adapts and Microch&rus, both of which occur in the English beds as well as in the phosphorites ; these being the last European representa- tives of the group. The Insectivora include Necrogymnurus (P.S.), allied to the Malayan Gymnura, Amphidozotherium, together with the existing genera Sorex and Talpa. More remarkable is the occurrence in the phosphorites of an insectivore described as Pseudorhynchocyon, which is believed to be a member of the family of jumping shrews (Macroscelidida), now confined to the Ethiopian region. The true Carnivora are represented by Eusmilus (P.), a highly specialised ally of the sabre-toothed tigers, as well as by the more cat-like sElurictis, and the generalised Pseudcelurus. In addition to species of true civets, referred to the living genus Viverra, the Viverrida include Amphictis (P.), Stenoplesictis (P.), and Palaoprionodon (P.); the two latter being generalised forms closely connecting the family with the Mustelidce, which is represented by Plesictis (P.). To the Canida may be assigned the genera Cynodon (P.S.), Cephalogale (P.), and Cynodictis, together with a species which may be included in Amphicyon 1 See table on p. 117. 192 EASTERN ARCTOG^A. [CHAP. (P.); while the creodont division of the order is represented by Hyanodon, Pterodon, Oxycena (P.), and Proviverra. Among the rodents we may note the squirrel-like Sciuroides (P.S.) and Pseudo- sciurus (S.), the existing genus Sa'urus, and the extinct Plesiarctomys and Plesiospermophilus, which likewise belong to the same family. Among the Muridce, Cricetus includes ancestral types of the hamsters ; while the dormice are represented by the existing genus Myoxus., As noticed previously, Theridomys, Nesocerodon, and Protechinomys seem to be ancestral forms allied to some of the existing South American rodents. The ungulates are very strongly represented; the pig-like group including Cebochoerus, Cheer opotamus and Elotherium (P.) among the Chcsropotamida^ Acotherulum, and Anthracotherium (P.) and Ancodus in the Anthracotheriidce. The anoplotheroids com- prise Anoplotherium, D aery t her ium, and Xiphodon ; Dichobunus and Ccenotherium (P.) are the characteristic forms among the Ccenotheriidce; and Dichodon, Gelocus, and Lophiomeryx among the DichodontidcR\ while the chevrotains are represented by Prodremo- therium (P.) and Bachitherium (P.). In the perissodactyle section of the same order, we have Pachynolophus (P.) to represent the Lophiodontidce, Palceotherium and Anchilophus in the Palceo- theriida, Protapirus (P.) as an ancestral form of the tapirs, and Rhinoceros (P.), Cadurcotherium (P.), and Hyrachyus (P.) as representatives of the rhinoceroses. The aberrant Chalicotherium has also one species from the phosphorites. The Effodientia in- clude Leptomanis (P.), Necromanis (P.), and Palceorycteropus (P.); while the existing genus Didelphys alone represents the mar- supials. It will be seen that in this fauna the existing generic types are very few, and if the whole of the extinct ones had been given, their relative proportion would have been still less. The ungulates were abundant, and among these the perissodactyles proportionately more numerous than at the present day ; while the anoplotheres are in some respect transitional between the latter and the typi- cal artiodactyles. All the ungulates had brachydont teeth, and annectant types between the modern pigs and ruminants were abundant ; the traguloids being the highest development among the artiodactyle section of the order. Creodont carnivores still V.] OLIGOCENE FAUNA. 193 persisted, although more modern types had already made their appearance on the scene ; and opossums flourished. The middle stage of the Oligocene is represented in Europe by the freshwater marls and clays of Hempsted in the Isle of Wight and the corresponding beds of Ronzon, near Puy-en-Velay, the lignitiferous strata of Cadibona in Liguria, the deposits of Fontainebleau and Ferte-Alais in France, and likewise by certain beds in Hungary and at Monte Promina in Dalmatia. Among the small fauna of this stage we may notice the following. In the Insectivora, Tetracus, an ally of the hedgehogs; Cynodon, Amphi- cynodon, Plesictis, and Hycznodon among the carnivores; Anthra- cotherium, Ancodus, Elotherium, Ccenotherium, Gelocus, and Rhino- ceros in the ungulates; and opossums (Didelphys). While this fauna is closely related to the preceding, it has lost a number of early ungulate types, such as Anoplotherium and Xiphodon among the artiodactyles, and Pal&otherium and Anchilophus in the perissodactyles. On the other hand, the pig-like forms, such as Ancodus, Anthracotherium, and Elotherium, attained an extra- ordinary degree of development. Among the creodont carnivores, we may note the final disappearance of the genera Pterodon and Proviverra, although Hycenodon still survived. The upper Oligocene (lower Miocene) fauna is a large and characteristic one, well represented in the freshwater beds of St Gerand le Puy, in the Allier, as well as in those of Weisenau and other localities in the neighbourhood of Mayence. Among the mammals may be mentioned the following ; existing genera being denoted by the prefix of a f. INSECTIVORA. fTalpa. fSorex. Geotrypus. Dimylus. t Myogale. Palseoerinaceus. Plesiosorex. t Erinaceus. CARNIVORA. Cephalogale. fViverra. Amphicyon. t Herpestes. Plesictis. Proaelurus. Potamotherium. Hyaenodon. Amphictis. L. 13 194 EASTERN ARCTOGyEA. [CHAP. RODENTIA. Theridomys. fSpermophilus. Archaeomys. f Sciurus. Issiodoromys. Chalicomys. t Myoxus. Titanomys. t Cricetus. UNGULATA. Anthracotherium. Csenotherium. Hyotherium. f Tapirus. Amphitragulus. t Rhinoceros. Dremotherium. MARSUPIALIA. f Didelphys. Of this fauna, Professor von Zittel writes that it seems at first sight closely akin to those of the middle and lower Oligocene ; the same ordinal and subordinal groups, and in many instances the same genera characterising the whole three horizons. In the lack of lemuroids, the reduced number or final disappearance of opossums, creodonts, and anoplotherioids, in the greater abundance of forms like Anthracotherium, Hyotherium, and Dremotherium, which were but poorly represented in the lower Oligocene, and in the number of new types, such as Tapirus, Amphitragulus (an ancestral chevrotain), Chalicomys (an early beaver), Titanomys (an ally of the picas), Erinaceus, Dimylus (a form connecting the shrews with the hedgehogs), Potamotherium (a generalised otter), Herpestes (mungooses), Procelurus (a primitive type of cat), we notice, however, the marked difference of this fauna from its fore- runners. Among the incoming genera it is noteworthy that there is none for which an ancestral type cannot be found in the lower Oligocene ; the main difference occurring in the more specialised characters of the members of the later fauna. With the exception of certain bats, insectivores, rodents, and the opossums (such as Vespertilio, Erinaceus, Sorex, Myogale, Talpa, Sciurus, Spermo- philus, Cricetus, Myoxus, and Didelphys], the majority of the genera are, however, still extinct. It is probable that the beds in the Balkans which have yielded remains of the North American Tertiary genus Titanotherium belong to some portion of the Oligocene epoch. We now come to the Miocene epoch, which, as at present re- stricted, forms in Europe but a small section of the Tertiary era. V.] MIOCENE FAUNA. 195 It includes the well-known freshwater strata of Sansan in Gers (the middle Miocene of the older geological classifica- tions), together with the corresponding beds of F^n°acene Steinheim in Styria, and likewise the somewhat newer (upper Miocene) deposits of CEningen, in Baden. Grive-St- Alban, in the valley of the Rhone, is likewise another well-known locality where mammaliferous strata of this age are developed; and, among other places, we may also mention Monte Bamboli in Italy, San Isidro in Spain, and Oran in Algeria. For the first time in Europe we meet with remains of true Primates, of which there are three genera belonging to the Simiidce, two of which, Dryopithecus and Oreopithecus *, are extinct, but the third seems scarcely separable from the existing Oriental Hylobates. In the Insectivora we meet with the existing European genera Talpa, Myogale (desmans), Erinaceus, Soreoc, and Crocidura ; while the extinct Lanth another turn seems to be allied to the tree-shrews ( Tupaid] of the Oriental region, and Galerix intermediate between the latter and the j Limping-shrews (Macroscelidtdce) of Ethiopian Africa. Among the Carnivora, where the creodonts have disap- peared, the cats are represented by the sabre-toothed tigers (Machcerodus] and Pseudcelurus. In addition to the existing genera Viverra and Herpestes, we have among the civet tribe the extinct Progemtta. The dogs include the existing Cam's, together with the extinct Hemicyon and Pseudocyon ; while the larger forms described as Dinocyon and Hycznarctus connect the former with the bears. The Mustelidce. are represented by species of the typical living genus Mustela, together with certain more or less closely allied extinct types ; and Enhydriodon filled the place of the modern otters. From among the rodents the generalised types allied to those now characteristic of Neogsea have all disappeared, nearly all the recorded forms apparently pertaining to existing genera. In the Sciuridce not only have we true squirrels (Sdurus\ but the Ethio- pian spiny squirrels (Xerus) are likewise represented, as are also the more widely distributed flying-squirrels of the genus Stiuro- pterus, which now inhabit both Eastern and Western Arctogaea 1 Some of the characters of these genera have been already mentioned on pages 1 80, 18 r. 13—2 196 EASTERN ARCTOG^A. [CHAP. Chalicomys, Cricetus, and Myoxus are survivors from the Oligocene; but porcupines (Hystrix] are new comers. Picas of the existing genus Lagomys are likewise to the fore ; and it is a question whether those distinguished by the name of Myolagus might not be included under the same title. A marked approximation to the modern type is likewise the characteristic feature of the ungulates of the European Miocene ; although in this group living genera still remain in the minority. The pigs (Sutda) include, for instance, the genus Hyotherium, in which the molar teeth are tuberculated and of the general type of those of some of the living members of the family ; and also the more aberrant Listriodon, characterised by the presence of a pair of transverse ridges on each of the teeth of the same series. A species of the existing West African genus Dorcatherium alone represents the chevrotains (Tragulidte) ; while we have forerunners of the deer (Cervidce) in the extinct Pal&omeryx and Dicroceros, both characterised by the simple structure of their antlers ; and Protragoceros — a generalised type of antelope — marks the first appearance of the hollow-horned ruminants (Bovida], which now form such a numerous and characteristic group in the fauna of Eastern Arctogaea. Perissodactyle ungulates are less numerous. Anchitheriiim, to some of whose distinctive characters allusion has been made above, constitutes the representative of the equine line at this stage ; and tapirs and rhinoceroses belonging to the existing genera were likewise common. Some of the latter were, however, still hornless, and in none was more than a single horn developed. The aberrant ChalicotheriidcB^ forming the last family of this section of the order, and characterised by the extraordinary resemblance presented by their claws and toes to those of eden- tates, are here represented by the gigantic Macrotherium. Finally the Miocene is notable as being the stage at which proboscideans first made their appearance on the scene in Europe. In this group we have species of Mastodon, which, as already explained, includes the ancestors of the modern elephants ; and likewise one of the more aberrant Dinotherium. Compared with the Oligocene, the loss of so many antiquated types, coupled with the appearance of proboscideans and man-like apes, and the general modern facies of all the mammals of the: V.] LOWER PLIOCENE FAUNAS. 197 Miocene, indicates the lapse of a considerable interval of time between the deposition of the two series of strata. And that this is really the case, is demonstrated by the fact that there occurs between the two a considerable thickness of marine deposits which have not hitherto yielded remains of land mammals. It may be noticed that while many of the insectivores and rodents from this horizon belong to genera now inhabiting the Eastern Holarctic region, among other forms we have marked instances of Oriental (Lanihanotherium. Hylobates] or Ethiopian ( Galerix, Dorcatherium, and Xerus) affinities in this assemblage; and it is thus evident that at the epoch in question there was no trace of the differentiation of Eastern Arctogaea into regions. Still more markedly are the same features displayed by the older Pliocene fauna of Europe and Southern Asia. This fauna, which was formerly regarded as of upper cen^Faun'a Miocene age until shewn by Dr Blanford to be unquestionably referable to the succeeding era of geological his- tory, had a very wide distribution ; and it is represented at certain localities, mostly at long distances from one another, by an extra- ordinary profusion of remains. One of these charnel-holes occurs at the village of Pikermi, near Athens, a second in the Isle of Samos, in the Turkish Archipelago, and a third at Mont Leberon, in Provence. This fauna is also met with locally in the valley of the Rhone, at the foot of the Pyrenees, in Spain, Asia Minor, and at Maraga in Persia. It is likewise represented in the regions lying to the north of the Alps, only here the number of forms is less, and the antelopes and giraffe-like ruminants, fitted for roaming over the open plains of the south, are conspicuous for their absence; their place being taken by forest-haunting deer. The sand-beds of Eppelsheim in Hessen-Darmstadt, together with strata in the neighbourhood of Vienna, and others in Hungary and Rumania, may be cited as localities where the northern section of this fauna is preserved. Taking first the European and Western Asiatic portion of this fauna, and leaving its Oriental members for subse- quent consideration, we find the Primates represented solely by an extinct genus of monkey, taking its name of Mesopithecus on account of presenting certain features 198 EASTERN ARCTOG^A. [CHAP. intermediate between the existing Semnopithecus and Macacus. The insectivores are likewise known only by a solitary form, a shrew (Sorex) ; but this is probably due to the nature of the strata being unfitted for the preservation of the remains of such small creatures. The Carnivora, on the other hand, were abundant, the Felidce being represented not only by the sabre-tooths (Machter- odus), but true cats (Felis) likewise making their appearance on the scene. Hyaenas display a great variety of development, there being one species of the typical genus Ifycena, with certain resemblances to the existing Cape form, while the more generalised types known as Lycyana and Hycenictis were likewise present, as were also species of Ictitherium^ and the allied Palhyana, which, as already mentioned, formed a connecting link between the hyaenas and the civets. True dogs seem to have been absent from this assemblage ; but Amphicyon still survived from the Miocene, and an aberrant form known as Simocyon made its appearance. Hycznarctus was likewise another survivor from the Miocene, and may be regarded as a forerunner of the true bears. Finally, in the weasel tribe (MustelidcB) we have representatives of the existing genus Mustela, as well as the extinct Pal&omephitis and Promeles, the latter being an ancestral type of the badgers. Rodents make but a poor show, as we have only the extinct beaver-like Chalicomys, a species of porcupine (ffystrix), and a representative of the curious little spiny mice (Acomys\ now characteristic of Syria, Palestine, and north-eastern Africa. A remarkable advance over their Miocene forerunners is displayed by the ungulates, especially those from Pikermi and Maraga. Here, in the artiodactyle section, we meet for the first time with true pigs of the genus Sus, which at this period ranged over the greater portion of Europe, and some of which attained very large dimensions. Water-chevrotains (Dorcatherium) serve to connect the Miocene representatives of their genus with the existing West African form; while muntjacs (Cervulus), now confined to the Oriental region, filled the place of the stags. Giraffe-like creatures were numerous, for not only have we true giraffes belonging to the existing Ethiopian genus Giraffa, but the gigantic hornless Helladotherium stalked over the plains of Greece, and the allied but horned Samotherium inhabited the area now LOWER PLIOCENE FAUNAS. 199 occupied by the Turkish Archipelago, and extended eastwards as far as Persia; Palceotragus being a smaller but allied form. The Bovidce are represented by antelopes, most of which present a marked Ethiopian facies, although Tragoceros (probably the direct descen- dant of the Miocene Protragoceros} is an aberrant form, with compressed horn-cores like those of the goats. And it may be remarked that most of the Pikermi antelopes have short-crowned molar teeth, in which respect they resemble the existing eland, kudu, and their allies. Of the Pliocene. forms, Palczorias, which is common to Southern Europe and Algeria, seems to be inter- FlG. 45. SKULL OF Palczorias. mediate between the kudus (Strepsiceros) and elands (Orias}\ while the so-called Protragelaphus is so closely allied to the existing Ethiopian harnessed antelopes (Tragelaphus) as to be included by some in the same genus. On the other hand, Palaoryx is nearly related to the gemsbok and its allies (Oryx), although with certain resemblances to the sable antelope group (Hippotragus). Gazelles (Gazella}, which are essentially inhabi- 2OO EASTERN ARCTOG^A. [CHAP. tants of open plains, were likewise abundant ; one being considered a near relative of the South African springbok. The genus Helicophora, on the other hand, closely resembles the water-buck group (Cobus), which is exclusively Ethiopian. In the perisso- dactyle division, the three-toed horses (Hipparion) seem to have approximated in general structure to the Ethiopian zebras, and, like those animals, may have been ornamented with dark and light stripes. While some of the Pliocene rhinoceroses were hornless, another was a two-horned species closely allied to the common African Rhinoceros bicornis, of which it may be regarded as the parent form. There is also an extinct genus (Leptodon), of some- what uncertain affinity ; while tapirs are found in the Eppelsheim beds, although not apparently in the southern area. The Chalico- theriidft were represented by the typical genus Chalicotherium (Ancylotheriuni), which, as we have seen, was a near ally of the Miocene Ma cr other him, and also occurs in the Oligocene phos- phorites. As in the Miocene, the proboscideans include only Mastodon and Dinotherium; the one species of the former ranging from Greece to Persia, but being different from all the Indian forms of the same epoch. Finally, the occurrence of an aard-vark (Orycteropus] both in Samos and Persia serves to accentuate the Ethiopian affinities of the southern section of this fauna. We have thus evidence that one and the same fauna extended from Spain and Algeria across Southern Europe to Asia Minor and Persia; and we may infer from the deposits at Samos, that what is now the y-Egean sea formed a tract of land connecting Greece with Turkey. It is further evident that there must have been free communication across the Mediterranean basin (which in Cretaceous times is known to have been a mare clausnm, in the physical, and not the political sense of the term) between Europe and Africa. This communication may have existed both by way of Gibraltar, and also between Italy, Sicily, and Malta on the one hand, and Tunis on the other; since the Plistocene mammals of the islands in question clearly indicate continental connection. While the antelopes and hipparions of this fauna prove the exist- ence of open plains during the lower Pliocene epoch, the host of individuals of Mesopithecus as unmistakeably point to the presence of extensive forest-tracts. In the northern section of the fauna, as V.] SIWALIK FAUNA. 2OI displayed at Eppelsheim, the Ethiopian affinities are much less apparent, aard-varks and the whole of the giraffe-group being absent, while tapirs and deer were abundant. That there was a more or less marked separation between the two areas thus seems evident; and the tapirs and muntjac-like deer, both of which seem wanting in the Siwalik fauna, are indicative, so far as they go, of Malayan affinities. Nearly related to that of Pikermi, Samos, and Persia, the celebrated Siwalik fauna of India and the adjacent countries presents certain well-marked differences; Fa^hk this being specially shown by the occurrence of several essentially modern types quite unknown in the former. More- over, there are a considerable number of peculiar genera which do not occur in the western fauna; while we also come across certain Miocene, and even Oligocene types, which are equally strange to the latter. Although in some cases these occur in beds which are not improbably of upper Miocene age, in others they appear mingled with the later forms ; but, in any case, they indicate a survival in this area of archaic types which at that time had com- pletely disappeared from Europe. Originally discovered in the outer ranges of the typical Hima- layan area, the Siwalik fauna has been traced towards the north-west into the Punjab, Kach, Sind, and the north-eastern frontier of Baluchistan ; the beds from the two latter areas being lower in the series than those from the typical Siwalik hills, and containing an older assemblage of forms, although several are common to all. An outlier of the same fauna occurs in Perim Island in the gulf of Cambay. Eastwards the Siwalik fauna ranged through Sylhet and Assam to Burma, whence it has been traced at intervals, as in Java, Sumatra, and the Philippines, into China and Japan. In China it extended from Yunnan in the south-west northwards through Szechuen to Kansu, and thence eastwards through Shensi to Shansi, its extreme eastern limit being indicated by the discovery of a Siwalik elephant's tooth at Shanghai. Northward of Kansu the fauna ranged into Mongolia, probably by way of the gap formed by the course of the Hwang-ho through the Ala-shan mountains — if such mountains existed at the time. And it is not a little remarkable that of the few Mongolian 202 EASTERN ARCTOG^A. [CHAP. forms at present known, two (Hycena macrostoma and Equus sivalensis) are identical with species from the Siwalik Hills1. As regards the fauna itself, we find, in the first place, the Primates much more fully represented than at Pikermi, and all by existing generic types. Of the man-like apes (Simiid&\ there is a chimpanzee (Anthropopithecus) presenting a more human type of dentition than its living Ethiopian cousins; while a single tusk indicates the former existence of an orang (Simla) allied to the living Bornean and Sumatran species. The other three generic types belong to the Cercopithecida, and include baboons of the Ethiopian genus Papio ( Cynocephalus), together with species of Semnopithecus and Macacus, the former genus being exclusively, and the latter mainly, Oriental at the present day, although both occur in the later Pliocene of Europe. Doubtless owing to the unsuitability of the strata for the preservation of small specimens, no remains of insectivores have hitherto been obtained. The Carnivora are, on the other hand, well represented; the Felicia including large and small species of the typical genus Felis, and apparently one of the allied Cynahirus (hunting-leopard), now exclusively Oriental and Ethiopian. Mach&rodus had two species ; and another form has been identified with the European Oligocene genus SEhiridis. Civets include species of Viverra larger than any now existing; this genus being also one now confined to the Ethiopian and Oriental regions, although more abundant in the latter than in the former. The CanidcE, in addition to a survivor of the Miocene Amphicyon, were represented by wolves and jackals (Cam's), as well as by a species apparently allied to the long-eared fox (Otocyon} of Africa. While in the Ursidce the generalised Hycenarctus still survived, true bears (Ursus) make their appearance for the first time, the single known Siwalik species presenting, however, a marked approximation in the characters of its skull and dentition to the Indian sloth-bear (Melursus). Among the few known representatives of the Mustelida, we have a large marten (Mustela), probably allied to the living yellow-throated Indian species ; a ratel, belonging to a genus (Mellivora) now restricted to India and Africa; and likewise an otter (Lutra) whose nearest affinities are with an existing Sumatran species. The same family also 1 Lydekker, Rec, Geol. Sw~v. India, Vol. xxiv. pp. 207 — 211 (1891). V.] SIWALIK FAUNA. 203 includes a member of the otter-like genus Enhydriodon, the other species being from the Italian Miocene. Among the most remarkable features of the Siwalik Carnivora is the survival of a species of Hycenodon, of which the remains have been discovered in the Punjab. The Rodentia are but very imperfectly known. They include a representative of the bamboo-rats (Rhizomys), which are now exclusively Oriental, and belong to the family Spalacida; and, among the Muridce, a species of Nesocia, which genus is likewise confined to the Oriental region. The other forms are a porcupine (Hystrix] and a hare (Lepus). A very long list is presented by the ungulates, which are numerous not only in generic, but likewise in specific types. The pig-like artiodactyles include, among the family Suidce. several representatives of the true pigs (Sus), some of which attained gigantic dimensions, while others are remarkable for the complex structure of their molar teeth, which show a marked resemblance to those of the existing Ethiopian wart-hogs (Phacochczrus). A still more elaborate structure is displayed by the corresponding teeth of the allied genus Hippohyus, which is peculiar to this fauna ; and the family is also represented by species of the European Miocene genera Hyotherium and Listriodon, the remains of the two latter being mostly obtained from the Punjab and districts to the west. The same areas are mainly those which have yielded remains of Anthracotheriidcz, although some of these have been discovered in Sylhet. In this family we have species of the European genera Anthracotherium and Ancodus, the former of FlG. 46. RIGHT UPPER MOLAR OF A SMALL SPECIES OF Merycopotamns. which is elsewhere unknown above the Middle Oligocene; and there are also three peculiar types, respectively known as Meryco- potamus, Hemimeryx, and Chctromeryx, differing from all the rest 204 EASTERN ARCTOG^A. [CHAP. in having only four columns on the crowns of the molars, as shown in the annexed figure, and thus presenting a marked approximation to the ruminants. The earlier Tertiary Chczropo- tamidcz likewise had a survivor in the genus Tetraconodon, which was represented by a large pig-like creature remarkable for the enormous size of its simple conical premolar teeth. The pig- like group closes with Hippopotamus, which makes its appearance on the scene for the first time in this formation, where it is represented by a generalised species with three pairs of incisor teeth in each jaw. Turning to the groups with fully-developed selenodont molars, we have first to notice the occurrence of fossil camels of the existing genus Camelus, which are unknown else- where except in the Algerian Plistocene. As we have seen that the Camelidce. were originally a New World group, it is interesting to note that these earliest Old World representatives occur in Asia instead of Europe; and it is further noteworthy that in the structure of their molar teeth the Siwalik camels retain evidences of affinity with the South American guanacos and vicunas which are lost in their living descendants. The Tragulidce contain repre- sentatives of the true chevrotains ( Tragulus) and water-chevrotains (Dorcatherium\ now respectively characteristic of the Oriental and Ethiopian regions, while among the deer ( Cervidce) we have species of the Oligocene European genus Palaomeryx, together with others belonging to Cervus, the representatives of the latter being all closely allied to existing Oriental types. Not improbably also a musk-deer (Moschus) should be included among the Siwalik Cervidce. Among the Giraffidce, in addition to true giraffes (Giro/a), which are common to the Pikermi beds, and extended eastwards into China, we have the peculiar gigantic antlered types re- spectively known as Vishnutherium, Sivatherium, Hydaspotherium, and Bramatherium, of which the first seems common to the Siwaliks of Burma and the Punjab, while the second is confined to the more easterly Himalaya, the third to the Punjab, and the fourth to Perim Island. They include the most gigantic of all ruminants, Sivatherium almost rivalling an elephant in bulk. Not one of the least curious features in this marvellous fauna is that while, as we have seen, deer of Oriental types were abundant, antelopes closely allied to those now inhabiting the Ethiopian v.] SIWALIK FAUNA. 205 206 EASTERN ARCTOG^EA. [CHAP. region — where deer are totally absent — were likewise extraordinarily numerous. Of the African genera we have a species of Bubalis intermediate between the hartebeests and the blesbok, a member of the sable antelope group (Hippotragus\ a kudu (Strepsiceros), an eland (Orias), and probably a representative of the water-buck group (Cobus). On the other hand, Oriental forms are not wanting, as proved by the occurrence of a nilgai (Boselaphus), and probably of a four-horned antelope (Tetraceros] ; while the widely- spread gazelles (Gazelld) were likewise present. Goats and oxen for the first time made their appearance ; the former group being represented not only by species belonging to the typical Capra, but likewise to the shorter-horned genus Hemitragus, now confined to India and Arabia. The oxen (Bos) included members of all the existing groups, that is to say typical oxen, bison, buffalo, and smaller forms with upright triangular horns nearly allied to the anoa of Celebes. The perissodactyle ungulates, so numerous in the earlier Tertiary formations, have now become proportionately much fewer as compared with the artiodactyles. While typical forms of Hipparion were present, one species differs from the rest by the loss of the lateral toes, and thus resembles the modern horses (Equus\ which here make their appearance for the first time. Rhinoceroses include not only hornless forms, but likewise one species allied to the existing Oriental Rhinoceros unicornis and R. sondaicus, and a third as closely related to the African Burchell's rhinoceros (7?. simus]. In the same group Chalicotherium is a survivor from older formations. Finally, the proboscideans exhibit a development unparalleled in any other formation or epoch. Dinotherium appears for the last time in the Siwaliks of Perim Island, Kach, Sind, and the Punjab ; while the mastodons include a large number of species, some of which present such a close approximation to the so-called stegodont elephants (which, as already mentioned1, are peculiar to this fauna) as to render it impossible to draw any well-defined demarcation between the genera Mastodon and Elephas. Not only does the Siwalak fauna include the aforesaid stegodont, or 1 Stipra, p. 172. V.] S1WALIK FAUNA. 2O; transitional elephants, but likewise one which may well have been the ancestor of the species now inhabiting India. Eastwards these transitional elephants and mastodons have been traced into Java, Borneo, China, and Japan ; and, as stated in an earlier chapter, there can be no doubt that the modern elephants were evolved in this area. Although, as shown in the foregoing survey, the Siwalik fauna differs in certain respects from that of Pikermi, Samos, Leberon, etc., yet there can be no hesitation in regarding the whole lower Pliocene fauna of Europe, North Africa, Asia Minor, and South and East Asia as essentially one ; and consequently at this epoch there was no possibility of distinguishing between the Palaearctic and Oriental regions. Whence Ethiopian Africa had by this time received the forerunners of its present higher mammalian fauna, we have, unfortunately, no decisive evidence. Writing some years ago, Dr Blanford1 seems to suggest that the irruption of the modem African fauna was anterior to the Pliocene. After referring to certain peculiarities connected with the existing mammalian fauna of India and the Malayan area, he observes that " these cases of isolation probably indicate that the animals belong to an older fauna, now partly replaced by newer types, and that the older fauna was common to India and Africa. It is very probable that these animals are descended from the ancient tropical fauna of the early Tertiary times. But, so far as it is possible to judge, the process of variation would have caused a greater distinction between forms so widely separated and exposed to such different conditions, if the period of isolation were great ; and it is difficult to suppose that the lands inhabited by the ancestors of the Simiidce, Lemuridce, Tragulida, and Manidce of the Oriental and Ethiopian regions can have been separated prior to the early part of the Miocene period." This is perfectly true so far as it goes, but since, as we have seen, genera like Hippopotamus, Bos, Capra, Equus, and Elephas are unknown previous to the Siwalik epoch, and some of them at least were evolved at or about that time in the Indian area, it seems necessary to assume the existence of a free land communi- 1 Manual of Geology of India, ist ed. p. Ixviii. (1879). 208 EASTERN ARCTOG^A. [CHAP. cation between the Ethiopian and Oriental regions at least as late as the lower Pliocene epoch. With regard to where this connec- tion was situated, we may note, in the first place, that Dr Wallace : was of opinion that even the Pikermi fauna made its way into Africa chiefly through Syria, although a brief connection of Europe with Tunis is admitted. When the passage in question was written, little or nothing was, however, known as to the Pliocene fauna of Algeria. And although this undoubtedly indicates a western connection between Europe and Africa, yet even in the Pliocene the Sahara probably formed, as now, a barrier2 across which the fauna of northern Africa could not pass south. Accord- ingly, even the Pikermi fauna may have come round byway of Egypt. Be this as it may, it seems clear that the Siwalik fauna entered Africa by way of Syria or Arabia, or possibly by both. The most direct line of communication would be via the Gulfs of Oman and Aden ; and some indication that such a line of connection may have existed is afforded by the distribution of the goats of the genus Hemitragus. As already stated, fossil species of this genus occur in the Siwaliks of Perim Island and the Himalaya, while of the three existing forms, one is Himalayan, a second confined to the Nilgiri and certain other South Indian ranges, and the third inhabits Oman. So far as it goes, the evidence of these goats is strongly suggestive of the former existence of a land-bridge across the mouth of the Persian Gulf, as otherwise we should expect to find living species in Persia and other parts of western Asia. If the existence of such a bridge be admitted, we only require another across the narrow strait of Bab-el-Mandeb to give a free line of communication between India and Africa in this direction. Whether, however, the migration from India to Africa took place at the north or south end of the Red Sea, or at both ends, it is certain that the connecting land must have been of consider- able width, and suited to the passage of mammals of all kinds. In referring to the nature of the connection, Dr Wallace3 remarks that " we may now perhaps see the reason of the singular absence 1 Geographical D^strib^ttwn of Animals, Vol. I. p. 288. 2 The idea that there was a Tertiary sea in the Sahara is incorrect; see Blanford, Quart, Journ. GeoL Soc. Vol. XLVI. p. 90 (1890). 3 Op. cii. p. 291. V.] SIWALIK FAUNA. 209 from tropical Africa of deer and bears ; for these are both groups which live in fertile or well-wooded countries, whereas the line of immigration from Europe to Africa was probably always, as now, to a great extent a dry and desert tract, suited to antelopes and large felines, but almost impassable to deer and bears." The Siwalik chimpanzee, however, indicates most unmistakably that the communication by way of Arabia or Syria between the Ethio- pian and Oriental regions must have embraced a forest-area, and accordingly have been of considerable width. With regard to the question why so many genera which existed in India and southern Europe during the Pliocene should have disappeared from those areas to live on in Africa, all we can say is that it is quite evident that a southern migration of the fauna has certainly taken place, and that this was probably induced by the cold heralding the approach of the glacial period. Although we have few, if any, decisive physical evidences of a cold period in India, yet the existence of a goat (Hemitragus) nearly allied to a Himalayan species in the ranges of southern India seems to indi- cate that such must have occurred, as it would be quite impossible for the ancestral form to have crossed the intervening plains under present conditions of temperature. It is further noteworthy that many of the animals which have disappeared from India, such as chimpanzees, hippopotami, giraifes, water-chevrotains, and ostriches, are precisely those which are now restricted to very hot climates ; whereas the lion, tiger, rhinoceroses, elephants, and monkeys, which both now or during the Plistocene are known to be capable of existing in cold climates, have persisted. Leaving these exceedingly difficult questions, two other points may be noticed in connection with the Siwalik fauna. In the first place, since the Siwalik hills themselves form ranges of con- siderable height on its southern flank, it is evident that the Himalaya was much lower during the lower Pliocene epoch than it is at present ; Dr Blanford l stating that the movement which led to its elevation "has been distributed over the Tertiary and post-Tertiary period, and a great portion in post-Plistocene." This will account for the community between the lower Pliocene 1 See Geol. Mag. Decade 3, Vol. ix. p. 166, note (1892). L. 14 210 EASTERN ARCTOG^EA. [CHAP. fauna of the Himalayan area and Mongolia, the Himalaya at that epoch not forming, as now, an impassable barrier to the north of the Oriental region. The second point relates to the survival in the Siwalik fauna of archaic forms, which had disappeared at that date from Europe. This fact, especially since old types such as lemurs and gymnuras are even now met with in the Oriental region, lends support to the view advanced in an earlier chapter1 that marsupials may have lived on in south-eastern Asia long after they had completely disappeared from Europe. Our knowledge of the later Pliocene faunas of Eastern Arctogaea is mainly confined to Europe, where at cene Kaunas.0" this period the general distribution of land and sea was apparently very much the same as at the present day. Spain was, however, connected with Africa, as was probably also Italy by way of Sardinia and Malta. A portion of Italy was, however, submerged, while in Belgium, Holland, and the south-east of England the sea intruded upon what is now land ; but, on the other hand, Britain was joined to the Continent. Few mammaliferous deposits of this age have been preserved to us, but among these are the Crags of the east coast of England (which contain numerous fossils derived from earlier formations), the fresh-water beds of the Val d' Arno in Italy, as well as others in the Auvergne, in the Rhone valley, at Roussillon, and in the neighbourhood of Montpellier. The following genera are in- cluded in this fauna, those which are extinct having an asterisk prefixed. PRIMATES. Semnopithecus. * Dolichopithecus. Macacus. INSECTIVORA. Sorex (Shrews). CARNIVORA. *Machaerodus (Sabre-tooths). Felis (Cats). Viverra (Civets). Hyaena. Canis (Wolves and Foxes). * Hyaenarctus. 1 Supra, p. 57. v.] LATER PLIOCENE FAUNAS. 211 CARNIVORA (cont.). Ursus (Bears). ^Elurus (Cat-bears). Mustek (Martens and Weasels). Lutra (Otters). RODENTIA. Arctomys (Marmots). * Chalicomys. Castor (Beaver). *Trogontherium (Giant Beaver). Cricetus (Hamsters). Microtus (Voles). Mus (Rats and Mice). Hystrix (Porcupines). *Pellegrinia, *Myolagusj Lagomys J Lepus (Hares) Sus (Pigs). Hippopotamus. Cervus (Deer). Alces (Elk). Cervulus (Muntjacs). * Palaeoryx. Gazella (Gazelles). Bos (Oxen). Tapirus (Tapirs). Rhinoceros. Equus (Horses). *Hipparion — very rare. * Mastodon! UNGULATA. Elephas J (Elephants). In this list by far the greater number of the genera are living ones, and if we removed from it types like Hyana, Hippopotamus, Rhinoceros and Elephas, which were spread during the Pliocene and Plistocene epochs over the greater part of Eastern Arctogaea, its Ethiopian resemblances are by no means strongly marked. Although the larger forms (as in the succeeding Plistocene epoch) 14—2 212 EASTERN ARCTOG^A. [CHAP. V. include a considerable number of genera now mainly confined to tropical or subtropical countries, the rodent fauna exhibits a marked Palsearctic facies, thus indicating an approximation to the existing state of things. Among the extinct rodents, Pellegrinia, from the Sardinian Pliocene, belongs, however, to the Octodontida, and is probably allied to the existing African Ctenodactylus. Tro- gontherium is a gigantic extinct type of beaver, which also persisted into the Plistocene. The deer include northern types unknown in the lower Pliocene. One of the most remarkable features of this fauna is the occurrence of a large species of sElurus, — a genus represented elsewhere only by the cat-bear or panda (^E. fulgens) of the eastern Himalaya, which, although formerly regarded as the type of a family by itself, is now included in the American Procyonidce (raccoons). The fossil species has been hitherto detected only in the English Crag ; the genus may, however, be expected to occur in the Siwaliks, since it is quite clear that it must have been originally connected with the American representatives of the family by forms inhabiting Eastern Asia. With the end of the Pliocene epoch this brief survey of the Tertiary mammalian faunas of Eastern Arctogaea may be brought to a close, since the Plistocene mammals can be more con- veniently considered under the headings of the different regions of this great province. While throughout the Oligocene, Miocene, and lower Pliocene epochs no trace of the present zoological regions of this half of the Arctogseic realm is shown, when the Upper Pliocene is reached there are faint indications of the demarcation of the Eastern Holarctic. At the time of the Plisto- cene, as will be shown in a later chapter, the Eastern Holarctic, Oriental and Ethiopian regions appear to have assumed a still more marked distinction, although this is to a great extent obscured by the wide range even at that epoch of genera like Hippopotamus, Rhinoceros, Elephas, Macacus, etc. Moreover, several species which are now confined to one of the three regions in question had then a more extensive distribution, so that it is only during the recent epoch that the Holarctic, Oriental, and Ethiopian regions attained the full faunistic peculiarities by which they are now characterised. CHAPTER VI. THE MALAGASY REGION. Limits — Mammalian Fauna — Relations of Madagascar to the Mainland. INCLUDED by Drs Sclater and Wallace within the Ethiopian region, Madagascar and the adjacent groups of islands were referred to a region apart by Dr Blan- ford1; this separation being justified not only by the mammalian fauna, but likewise by many other groups of animals. To quote Dr Wallace, this region "comprises, besides Madagascar, the islands of Mauritius, Bourbon, and Rodriguez, the Seychelles, and Comoro Islands. Madagascar itself is an island of the first class, being a thousand miles long, and about two hundred and fifty miles in average width. It lies parallel to the coast of Africa, near the southern tropic, and is separated by 230 miles of sea from the nearest part of the continent, although a bank of soundings projecting from its western coast reduces this distance to about 160 miles. Madagascar is a mountainous island, and the greater part of the interior consists of open elevated plateaus ; but between these and the coast there intervene broad belts of luxuriant tropical forests." It is this forest-district which forms the home of most of its peculiar fauna. As regards geological structure, it [appears from the researches of Messrs Cortese and Baron that, roughly speaking, a line drawn from north to south so as to divide the island into two longitudinal halves, gives an area of granitic and volcanic rocks on the right or eastern side, and on the left or western side one of sedimentary deposits, containing beds belonging to the Jurassic, Cretaceous, Eocene and recent epochs. Blown sand occurs in abundance 1 Appendix, No. 8, p. 76. 214 THE MALAGASY REGION. [CHAP. around the coast, and numerous old lake-basins or marshes, some of very large dimensions, form receptacles where remains of the later faunas have been preserved. With the exception of the Comoro group, which contain a few species, the non-volant mam- malian fauna is confined to Madagascar, so that the other islands do not properly come within the province of the present work. It is, however, important to observe that the Seychelles differ from almost all oceanic islands in consisting largely of granitic and other crystalline rocks. In an island lying so close to the African continent as Mada- gascar, the natural assumption would be that, if it fa^ia"1" possessed a mammalian fauna at all, such fauna would be closely allied to that of the mainland. As a matter of fact, precisely the reverse is the case, and out of a total of fully 28 genera of non-volant mammals now or recently inhabiting the island, only three are common to Africa. This, however, is by no means all, for out of these three genera two {Hippopotamus and Sus] are such as have probably crossed the intervening channel, although at a time when it was narrower than at present, while it is quite possible that the third (Crocidura) may have been introduced by human agency. Even this, however, scarcely gives a true idea of the case. In the first place, not only are the peculiar genera unknown in Africa, but they are equally strange to all the other regions of the world. In the second place, these genera belong to groups which form only a very small por- tion of the existing mammalian fauna of the Ethiopian region. At the present day, as will be more fully indicated in the following chapter, Ethiopian Africa is especially characterised by its nume- rous antelopes, as well by giraffes, zebras, rhinoceroses, elephants, hippopotami, wart-hogs, bush-pigs, lions, leopards and various other large cats, baboons, anthropoid apes, aard-varks, and ostriches. But, with the exception of the aforesaid bush-pig and extinct hippopotamus, not a single representative of any one of these groups is found in Madagascar. In place of such animals, Madagascar is populated by a host of lemurs, so numerous that the number of their species considerably exceeds that of all the other non-volant mammalian inhabitants of the island. Civet- and mungoose-like species, all pertaining to peculiar genera, alone VI.] THE FAUNA. 21$ represent the numerous Garni vora of the mainland; the Insectivora, in addition to the aforesaid Crocidura, or musk-shrew, include only the peculiar family of the tenrecs (Centettdce), which is confined to the island, and a representative of the Ethiopian family Potamogalida ; while the rodents comprise five genera of the cosmopolitan mouse-family (Muridce], more or less closely allied to one another, but different from any found elsewhere. The following is a list of the genera of non-volant Malagasy mammals ; those which are extinct being indicated by an asterisk, and the names of all the groups peculiar to the island printed in italics. PRIMATES. — LEMUROIDEA. LEMURID^E. Chirogale (Mouse-lemurs) ; 4 species. £jieiroAd.le<*s Microcebus (Dwarf Lemurs) ; 5 species. *• Opolemur (Fat-tailed Lemurs) ; 2 species. ^ Lemur (True Lemurs) ; 8 species. Mixocebus (Hattock) ; i species. ?^ Hapalemur (Gentle Lemurs) ; 2 species. 'y ( \ Lepidokmur (Sportive Lemurs) ; 8 species. 7.' $c»~*V\ Avahis (Avahi) ; i species. ? Propithecus (Sifakas) ; 4 species.^7- /^m'(Endrina) ; i species. * ME GAL AD A PID^E. * Megaladapis (Giant Lemur); i species. K««^l»*< CHIROMYID&. t\ $J±fL Chiromys (Aye-aye) ; i species. INSECTIVORA. SORICID.E. Crocidura (Musk-shrews) ; i species. CENTETID^E. Centetes (Tenrec) ; i species. Hemicentetes \ 2 species. Ericulus (Hedgehog-tenrec) ; i species. Echinops-j i species. Microgale (Long- tailed tenrecs) ; 3 species. Oryzorictes (Rice-tenrecs) : 2 species. 2l6 THE MALAGASY REGION. [CHAP. INSECTIVORA (cent.). POTAMOGALID^E. Geogale; i species. CARNIVORA. VlVERRID^E. CryptoproctincR. Cryptoprocta -, i species. Viverrinae. Fossa ; i species. Herpestinse. Galidictis (Striped Mungooses) ; 2 species. Galidia (Ring-tailed Mungoose) ; i species. Hemigalidia (Brown-tailed Mungoose) ; i species. Euplerince. Eupleres (Small-toothed Mungoose) ; i species. RODENTIA. MURID/E. Hypogeomys ; i species. Nesomys ; 2 species. Brachytarsomys ; i species. Hallomys; i species. Eliurus; 2 species. UNGULATA. Sus (Potamochcerus) ; t species. HIPPOPOTAMID^:. Hippopotamus ; i species (extinct). Considering the fauna in more detail, it may be first mentioned that the lemurs (Lemuroidea) differ from the higher, or Anthropoid Primates by their generally lower grade of organisation, as well as by certain features of the skull and internal anatomy which need not be more fully noticed here. They all have fox-like, expression- less faces ; and, with the exception of the aye-aye and the Asiatic tarsiers, they are characterised by the innermost pair of upper incisor teeth being separated from one another in the middle line. VI.] LEMUROIDS. 217 At the present day lemuroids are represented elsewhere only in the Ethiopian and Oriental regions ; the African forms being more nearly allied to the Malagasy types than are those of Asia. As stated in an earlier chapter, the group was, however, well represented in the lower Oligocene of western Europe, where certain forms (Mtcroc/ioerus) distinctly approximate some of the living kinds, although differing in the conformation of the first lower premolar tooth, which in the existing Lemuridce, assumes FlG. 48. SKULL OF LEMUR. tic. upper canine ; k. lower canine ; pm. premolars ; m. molars. the form and function of a canine or tusk. In the latter family (of which the distribution is coextensive with that of the suborder) the first three genera in the foregoing list belong to a subfamily (Galagina) distinguished by the elongation of the bones of the tarsus, and represented by an allied genus (Galago) on the African mainland. The next four genera constitute the typical subfamily (Lemurincz\ which is absolutely confined to Madagascar and some of the islands of the Comoro group, and of which the ring-tailed lemur (Lemur cattd) is one of the most familiar examples in European menageries. All these lemurs, which have long, although non- prehensile tails, differ from the first subfamily by the normal structure of the bones of the ankle. The third subfamily (Indrisinetds, the proper orthography is Orias. 248 THE ETHIOPIAN REGION. [CHAP. them as the representatives of separate genera. The existing species is now confined to the Ethiopian region, but since its fossilised remains occur in deposits of Plistocene age in Algeria, Spain, and Sicily, it is evident that, like the spotted hyaena and lion, it formerly enjoyed a much more extensive range. With the exception that a single species is found in Syria, the small rodent-like ungulates, known as hyraces, which constitute not only a family (Procavtidce) but likewise a sub-order (Hyracoidea) by themselves, are especially characteristic of the Ethiopian region, where they are represented by a large number of species, more FIG. 59. CAPE HYRAX (Procavia capensis]. particularly in the southern portion of the continent. Although these animals closely resemble the rhinoceroses in the structure of their molar teeth, they differ markedly from all the perissodactyles in having the carpus constructed on the linear type1, and from all other living forms of the order in that their single pair of upper incisor teeth grow continuously throughout life, as in the rodents. They were formerly divided into at least two generic groups, but both the terrestrial and arboreal forms are now included in the single genus Procavia. Nothing definite is 1 See figure 12 on p. 78. VII.] UNGULATES. 249 known as to the past history of the group1; but it has been suggested (p. 85) that they may be allied to certain extinct South American ungulates. The list of peculiar Ethiopian mammals is brought to a close by the aard-varks (Orycteropodida), which although generally in- cluded in the Edentata have nothing to do with the typical South American representatives of that order, and are here, together with the pangolins, regarded as forming an ordinal group — Effodientia — by themselves. Only a single genus (Orycteropus] now exists, of which there are two living Ethiopian species, and there are extinct species in the Pliocene of Persia and Samos. A skull from the Plistocene of Madagascar has been described as Plesi- orycteropus, and another genus occurs in the French Oligocene. Not improbably some members of the family entered Africa and Madagascar with the ancestral lemuroids and civets, but the dis- covery of the Pliocene forms renders it probable that the existing genus is a later immigrant. Finally, it may be mentioned that among the more widely- spread genera a few species of mammals are either now common to the Ethiopian region and India, or were so during the Plistocene age. In the Felida the lion (Felts leo), leopard (F. pardus), jungle- cat (F. chaus], caracal (F. caracal), and hunting-leopard (Cynce- lurus jubatus] still range over the two areas ; fossilised remains of the first three of these also occurring in the European Plistocene deposits. On the other hand, the spotted hyaena (Hycena crocutd), which lived in Southern India (as well as in Europe) during the Plistocene era, is now restricted to Ethiopian Africa; and the same is the case with the giant pangolin (Mains gigantea) of West Africa, fossilised remains of which have been discovered, in company with those of the spotted hyaena, in a cavern in Madras. The following table shows the genera and family of mammals now more or less exclusively restricted to the Ethiopian region ; the names of such as are practically peculiar to this area being printed in italic type. 1 I am informed that a skull belonging to an extinct member of this group has been discovered in the Pliocene of Samos, but no description has been published. 250 THE ETHIOPIAN REGION. [CHAP. I. Primates. 1. ANTHROPOIDEA. SlMIID/E. Anthropopithecus. Equatorial Africa. Fossil in Indian Pliocene. Gorilla. W. African. CERCOPITHECID^E. Co lob us. Cercopithecus. Largely W. African. Cercocebus. W. African. Theropithecus. N. E. African. Papio. Fossil in Indian Plistocene and Pliocene. 2. LEMUROIDEA. LEMURID^E. Galago. Equatorial and E. African. Perodicticiis. W. African. II Insectivora. MACROSCELIDID&. Fossil in European Oligocene. Macroscelides (including Petrodromus}. One N. African species. Rhynchocyon. E. African. ERINACEID^:. Proechinus. W. African. Myosorex. POTAMOGALID^:. Elsewhere only in Madagascar. Potamogale. W. African. CHR YSOCHLORID&. Chrysochloris (including Chalcochloris). S. and E. African. III. Carnivora. FELID^E. Cynaelurus. Elsewhere only in India, the same species being common to the two regions. VII.] LIST OF THE FAUNA. 251 III. Carnivora (cont.}. VlVERRID^. Genetta. One species in South Holarctic region. Poiana. W. African. Nandinia. W. and E. African. Helogale. W. and E. African. Bdeogale. W. and E. African. Cynictis. S. African. Rhynchogale. E. African. . Crossarchus. Suricata. S. Africa. Proteles. S. and E. African. Lycaon. S. and E. African. Fossil in European Plistocene. f Otocyon. S. African. MUSTELID^E. Mellivora. One African and one Indian species ; and another from the Indian Pliocene. Ictonyx. Ranges into Egypt, and perhaps Asia Minor. Pcecilogale. S. African. Galeriscus. E. African. IV. Rodentia. ANOMALURIDM. Anomalurus. W. and E. African. Idinrus. W. African. SCIURID/E. Nannosciurus. Elsewhere in Malaysia. Xerus. Fossil in European Miocene. MYOXID/E. Graphiurus. Claviglis. W. African. The right of the one species to generic distinction is doubtful. 252 THE ETHIOPIAN REGION. [CHAP. IV. Rodentia (cont.\ Pachyuromys. Mystromys. S. African. Otomys. S. E. and W. African. Dasymys. S. African. Malacomys. W. African. Dendromys. \ Limacomys. These represent a peculiar Ethiopian Steatomys. sub-family — the Dendromyina. Lophuromys. j Trilophomys. N. E. African, and (?) S. Arabian. Deomys. W. African. Alone represents a sub-family. Cricetomys. Saccostomns. W. African1. SPALACID.E. Bathyergus. \ Constitute a sub-family — the Ba- Georychus. thyergina ; the last of the four Myoscalops. being confined to Somaliland, Heterocephalus. ) while the first is S. African. Pedetes. The sole representative of a sub-family. OCTODONTID^:. Elsewhere at the present time only in the Neogaeic realm and N. Africa. Pectinator. N. E. African. Petromys. S. African. Triaulacodus. W., S. and E. African. V. Ungulata. n Eur°Pean and Asiatic ARTIODACTYLA AK1. UACiYLA. Plistocene and Pliocene, and HIPPOPOTAMI DAI. 4 . ,, , -r. , also in Madagascar. Formerly Hippopotamus. in lower Egypt. Sus ; the Potamochczrine group, frequently regarded as a distinct genus, is peculiar to the Ethiopian and Malagasy regions. Phacochcerus. 1 If Arvicanthis be accepted as a genus it should come here. VII.] LIST OF THE FAUNA. 253 V. Ungulata (cont.\ TRAGULID^E. Dor cat her ium. W. African ; fossil in European Miocene, and also in the Pliocene of India. GIRAFFID/E. ( Fossil in lower Pliocene of Europe and GiraffaA Asia. Bubalis (including Damaliscus}. Ranges into Syria and Tunis ; fossil in Indian Pliocene. Connochcztes. Cephalophus. Madoqua. N. E. and E. African. Nanotragus. W. African. Nesotragus. Rhapiceros. Oribia. S. African. Oreotragus. S. and E. African. Dorcatragus. Somaliland. Pelea. S. African. Cobus. Fossil in Indian Pliocene. Cervicapra. S. W. and E. African. sEpyceros. S. and W. African. Ammodorcas. Somaliland. Lithocranius. E. African. Hippotragus. Fossil in Indian Pliocene. Oryx. Ranges into Syria. Tragelaphus. Perhaps fossil in European Pliocene. Strepsiceros. ) -, . f Fossil in Indian Pliocene. Onas. I PERISSODACTYLA. RHINOCEROTID^E. Rhinoceros ; the species without front teeth are now peculiar to the Ethiopian region, although allied forms occurred in the European and Asiatic Pliocene and Plistocene. EQUID/E. Equus ; all the striped species confined to this region. 254 THE ETHIOPIAN REGION. [CHAP. V. Ungulata (cont.). 3. HYRACOIDEA\ PROCAVIID/E. I Range into Syria. Procavia. VL Effbdientia. Ethiopian and Oriental. ORYCTEROPODIDJS. Fossil in French Oligocene. Orycteropus. Fossil in Pliocene of Samos and Persia. Among groups other than mammals, attention may be directed to the remarkable difference between the birds of Ethiopia and those of Madagascar. On this point Dr Blanford1 writes that "the most characteristic African families, such as plantain-eaters (Musophagidce), colies (Collidce), and wood- hoopoes (Irrisoridce), barbets, hornbills, secretary-birds (Serpen- tarius], and a number of genera, such as Lamprotornis (glossy starlings), Buphaga (ox-peckers), Laniarius, and Telephonus, that are the common and familiar birds of every part of Africa south of the Sahara, are entirely wanting in the Mascarene Islands, in- cluding the Seychelles, Mauritius, etc., while no fewer than four peculiar families and a number of genera confined to the archi- pelago replace them. Amongst the Mascarene birds, too, are found several representatives of Oriental genera, or genera closely allied to Oriental types, and without any near Ethiopian relations. Foremost among these are certain bulbuls, forming the genera Ixocincla and Tylas, the former composed of species which have been usually referred to the typically Oriental genus Hypsipetes, and the latter nearly affined. In fact, as was shown by Geoffrey St Hilaire, and as Hartlaub has since pointed out, there is in the Mascarene avifauna a more marked connexion with Indian than with Ethiopian types. In the Seychelles, especially, out of the seven Passerine genera represented by peculiar species, three, Nectarinia, Zosterops, and Tchitrea, are Indian and African, one, Foudia, is Ethiopian, but not Indian, and two, Copsychns and Hypsipetes, or Ixocincla, are Indian but not African." All this is confirmatory, not only of the right of Madagascar and the Mascarenes to form a region by themselves, but likewise 1 Appendix, No. 8, p. 89. In this quotation the English or Latin names have in some cases been added to the original. VII.] PAST HISTORY. 255 of -the distinction between the Ethiopian and Oriental regions, which some have proposed to unite. It is, however, somewhat remarkable that secretary-birds (Serpentarius) are unknown in Madagascar, seeing that they are represented in the upper Oligo- cene of France, and may therefore be presumed to have entered Ethiopia with the ancestral lemuroids and civets. Finally, the ostriches (Struthio), which are now mainly confined to Africa and Syria, are evidently recent immigrants into the region, the genus being represented in a fossil state in the Indian Pliocene. With the exception of the occurrence of remains of certain existing species, such as Rhinoceros simus, Phaco- charus, etc., in the superficial deposits of southern Africa, nothing is known of the mammalian Tertiary palaeontology of the Ethiopian region. Fortunately, however, the clue given by the existing fauna of Madagascar and the Tertiary faunas of Europe and southern Asia enables a considerable portion of the past history of the population of the region to be given with a fair degree of completeness. And here, with one important exception, Dr Wallace's explanation, as given in the Geographical Distribution of Animals1 > may be accepted almost in its entirety; — the one exception being, as mentioned in an earlier chapter, that the Sahara was never a sea during Tertiary times ; although it appears always to have formed a barrier between northern Africa and Ethiopia. As already mentioned, the an- cestral types of the existing mammalian fauna of Madagascar ob- tained an entrance into Ethiopia some time during the Oligocene period, and soon after ranged over the whole of what are now the Ethiopian and Malagasy regions, which were then united and possessed a common fauna. During the Pliocene age, when Madagascar had become isolated, came the great irruption into Ethiopia, of the higher and larger mammals, such as apes, monkeys, ungulates, etc., which were then flourishing all along southern Europe and Asia. Finding the country unoccupied and eminently suited to their existence, these rapidly attained a development now unequalled in any other part of the world ; many new genera being apparently evolved within the Ethiopian area, although a large 1 Vol. i. p. 285 — 292. 256 THE ETHIOPIAN REGION. [CHAP. number were already in existence at the time of the southern migration. Several of these existing genera are met with in the Pikermi deposits of Greece, but more were confined, at this epoch, to the Pliocene of Persia, Samos, and India ; and it may there- fore be assumed that the great migration was by way of Syria or Arabia. Dr Wallace has indeed expressed the opinion that a certain number of types — among them the elephants and rhinoce- roses— obtained an entrance to the westward of Tunis ; but there are no true elephants in the Pikermi deposits, and apparently none in those of Persia, whereas their remains abound in the Siwalik Hills. As to the rhinoceroses, although the Pikermi species is closely allied to the African Rhinoceros bicornis, the Siwalik R. platyrhinus is equally close to R. simus ; and in the Siwaliks we meet with chimpanzees (Anthropopithecus), baboons (Papio), ratels (Mellivora), hippopotami, water-chevrotains (Dorca- therium\ and several genera of Ethiopian antelopes, all of which are totally unknown in the Pikermi beds. Ostriches, too, are first known in the Siwaliks ; while aard-varks occur in the Persian and Samos beds. All the evidence accordingly points to the great immigration having taken place along the eastern side of the continent ; and the existence of certain species of mammals which are either still common to India and Africa, or which were so during the Plistocene epoch, lends support to this view. Further testimony in this direction is aiforded by the occurrence of closely- allied generic types in the Ethiopian and Oriental regions. Among the lemuroids, for instance, the Oriental lorises (Nycti- cebus and Loris) are replaced in Western Africa by the potto and awantibo (Perodicticus) ; while in the Viverrida the true linsangs (Linsanga) of the eastern half of the Oriental region are repre- sented in Fernando Po by the allied Poiana, and the Oriental palm-civets (Paradoxurus} have very close allies in the two species of the Ethiopian genus Nandinia. A less marked instance is afforded by the occurrence of the water-chevrotain (Dorcatheriuni} in West Africa and of the true chevrotains in southern India and the eastern half of the Oriental region. And here it may be remarked that especial stress has been laid upon the much greater resemblance that exists between the fauna of the eastern, or Malayan, division of the Oriental region and VII.] ETHIOPIAN AND MALAYAN FAUNAS. 257 Western Africa, than between that of peninsular India and Eastern and South Africa ; large man-like apes and linsangs being confined in the Oriental region to its eastern half, while palm- civets, lorises, and chevrotains are more abundant there than in other parts of the same region. This, however, appears to be mainly or entirely due to similarity of climatic conditions, and not to original distributional distinctions. And, it may be re- marked, increased acquaintance with the fauna of Ethiopia tends to show that types formerly thought to be confined to the western half of Africa really extend far to the eastward ; chimpanzees, for instance, being now known to range as far east as Ugogo, while the genus Nandinia, which was originally known solely by a West African species, is now proved to have an eastern representative in Nyasaland. The similarity between the fauna of the Malayan sub-region and that of Western Africa naturally leads on to the consideration of the former land-connection between India and Africa. Writing on this subject, Dr Wallace1 observes that "we may now perhaps see the reason of the singular absence from tropical Africa of deer and bears ; for these are both groups which live in fertile and well- wooded countries, whereas the line of immigration from Europe to Africa was probably always, as now, to a great extent a dry and desert tract, suited to antelopes and large felines, but almost impassable to deer and bears. We find, too, that whereas remains of antelopes and giraffes abound in the Miocene2 deposits of Greece, there were no deer (which are perhaps a somewhat later development), neither were there any bears, but numerous forms of Felidce, Viverrida, Mustdida, and ancestral forms of Hyczna, exactly suited to be the progenitors of the most prevalent types of modern African zoology." As mentioned in an earlier chapter, since this passage was written the discovery of the remains of a species of chimpanzee in the Indian Siwaliks has shown quite clearly that the line of communication between India and Africa must have included a wooded tract comparable to the existing equatorial African forest region ; and this would be true even if the migration 1 Op. dt. p. 291. 2 The Pikermi beds were formerly universally held to be of Miocene age. L. I 258 THE ETHIOPIAN REGION. [CHAP. had taken place from Africa to India, which was not the case. Evidence of such a tract is, I believe, afforded by the occurrence of fossilised tree-stems in many districts which are now desert. And along this tract there can be little doubt that the ancestors of the mammalian types now common to the West African and Malayan sub-regions originally wandered from their common Indian home. Subsequently the whole of the countries lying between eastern Africa and India have become deforested, while in Africa itself the forest-area has shrunk away from the eastern side of the continent. This leaves the question of the absence of bears and deer from Africa without any adequate explanation. Bears are, however, in the main, mountain animals, some of which, like the isabelline bear of the Himalaya, inhabit districts where there is but little forest ; and it is noteworthy that, with the exception of the sloth- bear, which forms a genus apart (Melursus)^ there are no bears in India proper, although a fossil species allied to the sloth-bear occurs in the Siwaliks. This being so, when we take into account the absence of ursine remains from the Pikermi and Persian beds, there is nothing very wonderful in the fact that none of these animals entered Ethiopia during the great Pliocene migration. The absence of all Cervidce is more difficult to explain, seeing that deer of an Oriental type are abundant in the Siwaliks, while they are also sparingly represented in the Pikerrni beds. Typical deer of the red-deer group are, however, totally wanting in the Siwaliks, as they are in the Oriental region at the present day ; and we are, therefore, perfectly able to account for their absence from the Ethiopian region, although they occur in Africa north of the Sahara. With regard to the absence of Oriental types of deer in Ethiopia, it can only be said that it is as difficult to see any reason why these should have continued to flourish since the Pliocene in the Oriental region without ever having entered Africa, as it is to explain why giraffes, hippopotami, and ostriches should have disappeared from the former area to survive in the latter. Another difficulty is presented by the case of the pigs, bat here it may be suggested that the absence of the typical group of the genus Sus from the whole of Ethiopia, with the exception of Sennaar, may perhaps be accounted for by all the other species of VII.] LEADING CHARACTERISTICS. 259 that genus which originally entered the country having been developed into the more specialised Phacochxrus. Attention has already been called to the fact that the molars of some of the Indian Tertiary species of Sus show a distinct approximation to those of Phacochairus, and further evolution might easily lead to the development of the latter. Since this genus is unknown in a fossil state from other regions, is it improbable that it has originated from Sus within the limits of its present habitat ? Summarising the results of the foregoing survey of the mam- malian fauna of Ethiopia, it would appear that the Sahara has for a very long period formed a barrier between Ethiopian Africa and the northern part of the continent. When first populated by Tertiary mammals, Ethiopia and Madagascar were in union, and formed but a single zoological province, which would seem to have been to a great extent isolated from the rest of the Old World during the Miocene epoch. Had such conditions persisted |his province would have been entitled to form a primary zoological realm by itself. During the Pliocene epoch, however, Madagascar became separated, while a more complete union of the continent with Asia by way of Syria or Arabia permitted the influx of larger and higher mammals from the eastward. Hence there is a most intimate relationship between the Pliocene fauna of India and the one now inhabiting Ethiopia ; but the distinc- tion between the two areas at the present day is fully sufficient to justify the separation of the Ethiopian from the Oriental region. Of all the zoological regions of the world, the Ethiopian may be regarded as the one which has been most recently evolved ; and it may be shortly characterised by the sole possession of the gorilla and chimpanzees ; the absence of bears and deer ; and the presence of the African elephant, hyraces, rhinoceroses devoid of front teeth, zebras, hippopotami, wart-hogs, giraffes, numerous genera of antelopes, and aard-varks, as well as by the great de- velopment of its large ungulates in general. It shares with the Oriental region the distinction of being the sole habitat at the present day of man-like apes, true civets ( Viverra), linsangs, palm- civets1, ratels (Mellivora), elephants, rhinoceroses, and pangolins 1 Also in the Austro-Malayan region. 17—2 260 THE ETHIOPIAN REGION. [CHAP. (Mam's), together with the rodent genera Nannosciurus, Golunda and Atherura. The probable relationship between Ethiopia and Neogaea has been sufficiently discussed in the third chapter. Although, from the mammalian point of view, Ethiopia is a very modern region, as a continent it is one of the oldest, the greater portion of its area having been land since the Palaeozoic epoch. As has been shown in an earlier chapter, in Palaeozoic times southern Africa formed a portion of the great southern or equatorial continent distinguished from more northern lands by the peculiar characteristics of its flora; and it is probable that it remained connected with India until late in the Secondary epoch1; as is proved by the identity of the flora and reptiles of the two areas. Early, however, in this epoch there must also have been free communication with Europe, as shewn by the close alliance of the anomodont reptiles of the two continents, and likewise by the occurrence in both of the mammalian genus Tritylodon. In the Cretaceous, so far as vertebrates are conf cerned, our knowledge of the Ethiopian region is practically a blank. In conformity with the plan adopted in other cases, the sub- regions into which Ethiopian Africa may be divided Sub-regions. . . will be treated very briefly. Writing of the desert-tracts of the Saharan sub-region, where the necessary conditions for existence are largely wanting, Dr Heilprin observes that " there is a marked impoverishment of the fauna. The more formidable carnivores, such as the lion and the leopard, are absent from most districts, leaving their places to be filled by some minor cats, the hyaena, jackal, fox, and fennec. The hoofed animals are represented (in some parts) by the buffalo, and a limited number of antelopes (Gazella, Oryx, and Addax). Among rodents the families of rats (Muridce) and jumping-mice (Dipodidcd) are fairly represented, in addition to which we have the porcupine and hare (Lepus mediterraneus}. The ostrich is sufficiently abundant throughout most of the sub-region." In marked contrast to the poverty of the Saharan districts, is the remarkable richness of the great equatorial forest-tract, 1 Vide siipra, p. 224. VII.] SUB-REGIONS. 26l which is the exclusive home of all the man-like apes and of the lemurs of the genus Perodicticus. To this sub-region also belong PotamogaUi the African linsang (Poiana), as well as several other genera of Viverridcz, such as Nandinia and Helogale. Among the rodents the flying-squirrels of the family Anomalurida are very characteristic of the forest tract, and the peculiar murine genus Deomys, as also Saccostomus, is restricted to it. The water-chevro- tain (Dorcatheriuni) is solely West African, as is the small Liberian hippopotamus ; while certain genera of antelopes, such as the duikers (Cephalophus), harnessed antelopes (Tragelaphns\ and elands (Orias), have larger and finer representatives here than elsewhere. The giant pangolin (Manis gigantea), the largest member of its genus, is likewise a West African form. Although South Africa has a certain number of mammalian genera, such as Cynictis, Suricata, Otocyon, Pcecilogale, Bathyergus, Pedetes, Petromys, and Pelea, peculiar to it, others, such as the golden moles (Chrysochloris) and aard-wolf (Proteles], have been proved to extend far up the east coast, the latter occurring in Somaliland. Hence it seems advisable to unite Dr Wallace's South African sub-region with that portion of his East Central sub-region which is not included in the equatorial forest-tract. Of this East Central sub-region, as it may be collectively called, it will suffice to say that it is the home of the greater number of the characteristic Ethiopian mammals exclusive of those restricted to the forest-tract. Here antelopes attain their greatest numerical development, both as regards genera and species; and here is also the true home of the lion and the spotted hyaena ; while to this sub-region are confined the Cape hunting-dog (Lycaon) and the aard-wolf (Proteles). The distribution of other genera is sufficiently indicated in the table already given. The light which has recently been thrown upon the mammals of northern Somaliland has shown that, as regards antelopes at least, these are so peculiar that it may be questioned whether this tract is not entitled to be separated as a sub-region by itself. According to the lists given by Dr Sclater1 and Capt. Swayne2, 1 Natural Science, vol. I. p. 264 (1892). 2 Seventeen Trips to Somaliland, London (1895). 262 THE ETHIOPIAN REGION. [CHAP. there are no less than sixteen northern Somaliland, of which the 1. Bubalis swaynei. 2. Madoqua swaynei. 3. „ phillipsi. 4. ,, guentheri. 5. Oreotragus saltator. 6. Dorcatragus megalotis. 7. Cobus ellipsiprymnus. 8. Gazella pelzelni. species of antelopes found in names are as follows, viz. : 9. Gazella spekei. 10. ,, soemmerringi. 11. Ammodorcas clarkei. 12. Lithocranius walleri. 13. Oryx beisa. 14. Tragelaphus decula. 15. Strepsiceros kudu. 1 6. imbubis. Among these Nos. i, 2, 3, 4, 6, 9, n are quite peculiar to this district, while No. 12, which, like Nos. 6 and n, is the sole repre- sentative of its genus, is only found elsewhere along the east coast as far south as the Tana river. Another generic form peculiar to Somaliland is Heterocephalus, including two small naked rodents with burrowing habits ; while in the same order the single species of Pectinator is restricted to this district. Among species may be mentioned two musk-shrews (Crocidura smithi and C. somalica), a hedgehog (Erinaceus sclateri), as well as a banded mungoose (Crossarchus somalicus). The Somali ostrich seems likewise to represent a species by itself. On the other hand, in addition to those mentioned in the foregoing list of antelopes, there are several East or South African mammals, such as the aard-wolf, which range into Somaliland, and further evidence is perhaps desirable before the right of that country to form a separate sub- region can be admitted. The case is more clear with regard to south-eastern Arabia, whence Mr O. Thomas1 records the following fifteen species of land-mammals, viz. : Xantharpyia amplexicaitdata, Taphozous nudi- ventriS) Rhinopoma microphyllum, Erinaceus niger, Crocidura murina, Herpestes albicauda, Cants pallipes, C. leucopus, Gerbillus dasyurus, Mus rattus, Lepus omanensis, Gazella muscatensis, Oryx beatrix, Hemitragus jayakari, and Procavia syriaca. Of these Mr Thomas remarks that their geographical relationships " are, as might be expected, about equally with Africa and India, three of 1 Proc. Zool. Soc. 1894, p. 449. In one case the generic title has been altered, in order to bring it into harmony with the system followed in this work. VII.] SUB-REGIONS. 263 the species being distinctly African in affinities, three Indian, and the remainder either peculiar or widely-spread, and of no special significance." The association of a goat belonging to the Oriental genus Hemitragus with such an essentially Ethiopian animal as a hyrax {Procavia}, shows that we are here truly on the border-land between the two regions in question. In conclusion, from whatever aspect it be regarded, Ethiopia is one of the most interesting of the regions of Arctogaea; and if, as is suggested in an earlier chapter, it has been the feeder by means of which South America received its earliest mammalian Tertiary fauna, it is entitled to an importance above all the other zoological regions of the realm to which it pertains. CHAPTER VIII. THE ORIENTAL REGION. Sub-regions — Characteristics of the Mammalian Fauna — Past History of the Region — Malayan sub-region — Nicobars, Mentawi, and Christmas Islands — Philippine sub-region. FAR inferior in extent to the Ethiopian, the Oriental region is taken to include those portions of the continent of Asia lying south of the Holarctic region (with the exception of southern Arabia, which is Ethiopian), together with the islands of Ceylon, Formosa, the Philippines, Sumatra, Java, Borneo, and numerous smaller ones. In India the northern limits of the region are formed by the higher ranges of the Himalaya, while " Wallace's line" constitutes the eastern boundary denning it from the Austro- Malayan region of the Notogaeic realm. In a region so diversified as the Oriental, it would not be natural to expect a homogeneous fauna; and, as a matter of fact, there are in this respect great diversities between the different portions of the region, many of the peculiar genera having a very restricted distribution. Never- theless, the positive and negative features of the mammalian fauna of the region as a whole are sufficient to indicate its zoological unity, and also to differentiate it from the Ethiopian region, to which it is now most closely allied. In the Himalaya there is a gradual transition towards the Holarctic fauna ; and it is probable that in this portion of the area the differentiation between the Oriental and Holarctic faunas has been largely due to the eleva- tion of the Himalaya itself, which has taken place entirely since the early part of the Tertiary period, and is to a considerable extent of Post-tertiary date. It has already been shown that the older Pliocene fauna of northern India and Burma contained a CHAP. VIII.] PHYSICAL FEATURES. 265 remarkable admixture of mammalian genera now respectively confined to the Ethiopian and Oriental regions, together with some of a Holarctic facies ; and the completion of the elevation of the Himalayan chain was probably an important factor in the dispersal and differentiation of this common fauna. Holarctic types are again met with in force in the open desert regions on the north-western frontier of India. On the other hand, the fauna of the Philippines exhibits an approximation to that of the Austro-Malayan region, and thus shows a blending between the Arctogaeic and Notogaeic realms. In physical features the Oriental region displays great variation, a large portion of peninsular India consisting of open dry grassy plains, whereas the slopes of the eastern Himalaya, together with the greater part of Assam, Burma, and the Malayan countries are clad with luxuriant forests ; these tropical or sub-tropical forest-regions being those where the fauna attains its fullest and richest development. The poorest part of the region is, as Dr Wallace1 observes, the great triangular plateau forming the Indian peninsula; this area differing remarkably from the Himalaya in its geological features, and having been land since an extremely ancient date, whereas the Himalayan area consists very largely of marine formations. Since it is stated in the passage cited that peninsular India during the Tertiary period existed as an island entirely discon- nected from the Himalaya and Burma, it may be well to quote the later and more authentic views of the authors of the Manual of the Geology of India2 on this subject. After reference to the extent of Eocene rocks in northern India, it is there stated that " the Peninsula of India in Eocene times was part of a tract of land, perhaps of a great continent united to Africa ; that there was a sea to the eastward, extending far to the north-east, in the region now occupied by the Assam hills, and another sea to the north-west, covering great part, if not the whole, of Persia, Baluch- istan, the Indus plain, and a portion of the upper Ganges plain. An arm of this sea extended from the north-west up the Indus valley in Ladak. The Himalaya, and perhaps Tibet, wholly or in part, were raised above the sea ; but formed in all probability 1 Geographical Distribution of Animals, vol. I. p. 314. 2 First edition, pt. I. p. liii. 266 THE ORIENTAL REGION. [CHAP. land of moderate elevation. Whether the Himalayan land was united to the Peninsula is, of course, uncertain — but very pro- bably it was ; for there is no evidence of marine conditions having existed in the Ganges plain to the east of the Dehra Dun ; and if the ferruginous bands of the Subathu group be laterite, as they appear to be, the trappean detritus composing them must have been derived, in all probability, from the peninsular area; and the latter must consequently have extended northward to the base of the Himalayas, in the neighbourhood of Umballa.... In Miocene times, although marine conditions prevailed throughout western Sind, the area of the sea was very much smaller than in the Eocene period ; for all the marine beds of the Punjab and Sub-Himalayas are destitute of marine fossils, and are probably fluviatile de- posits." From this it would appear probable that during the whole of later Tertiary times, at least, there has been no isolation of peninsular India from the eastern Himalaya and the Burmese countries ; and consequently that the differences between the faunas of these areas are mainly or solely due to their differences of physical features and climate. By Dr Wallace the Oriental region is divided into four sub- reerions ; namely (i) the Indian, comprising the whole Sub-regions. J ^ ' of upper India, (2) the Ceylonese, including southern India and Ceylon, (3) the Indo-Chinese, embracing Assam, Burma, and such portion of China as lies within the limits of the region, and (4) the Indo-Malayan, which includes not only the Malayan archipelago and islands, but likewise the Philippines. So far as India and its dependencies are concerned, the follow- ing amended scheme has been proposed by Dr Blanford l, viz. : i. Himalayan. The southern slopes of the Himalaya, from the base to about the limit of trees. ii. Indian. India from the base of the Himalaya to Cape Comorin, with the exception of the Malabar coast, but with the addition of northern Ceylon. iii. Malabar or Ceylonese. The Malabar coast and the neigh- bouring hills as far north as the Tapti river, together with southern Ceylon. 1 Fatina of British India — Mammalia, p. v. VIII.] SUB-REGIONS. 267 iv. Burmese. All Burma except south Tenasserim, and with the addition of Assam and the intervening countries, v. Malayan. South Tenasserim, the Malay Peninsula, and the Malayan Islands as far as Wallace's line. Whether the Philippine Islands should be included in this sub-region, or should form one by themselves, may be doubtful, vi. Indo-Chinese. Although not free from objection, this term may be employed for the sub-region indicated by that portion of China coming within the limits of the Oriental region. Regarding these sub-regions in general, Dr Blanford observes that some " may require further subdivision. Thus the fauna of the North-west Provinces and Punjab differs considerably from that of southern India, and both areas exhibit zoological dis- tinctions from the forest-clad tracts of south-western Bengal. There is also much difference between the animals of Pegu and Arakan, on the one hand, and those of the drier regions of upper Burma on the other ; and even greater distinctions may be traced between those found in the sub-tropical and those inhabiting the temperate regions of the Himalaya. On the other hand, the sub- tropical Himalayas were united with the Burmese sub-region by Wallace, and the two are, perhaps, zoologically more allied to each other than to any other sub-region." Recent discoveries clearly indicate that the Philippine Islands, exclusive of Palawan and the Calamianes, should form a sub-region by themselves. Taking the Oriental region as a whole, it may be stated that the number of peculiar generic types of mammals is less than in the case of the Ethiopian; and that there are but two families absolutely confined to it, although a third is very nearly so. While sharing with the Ethiopian region the want of several groups of insectivores and rodents, such as the typical shrews (Sorex), marmots (Arcto- mys), and voles (Microtus), it lacks some of the other deficiencies of that region, true pigs (Sus) and deer being abundant, although the latter belong to groups distinct from those of the Holarctic region, while bears, belonging to two genera, are likewise met 268 THE ORIENTAL REGION. [CHAP. with. Wart-hogs (Phacochcerus), aard-varks (Orycteropus\ hyraces (Procaviid(B\ and jumping-shrews (Macroscelididce) are among some of the more characteristic Ethiopian animals which are wanting in the Oriental region at the present day, and have not hitherto been obtained there in a fossil state. Giraffes, a number of genera of antelopes, and hippopotami, are equally conspicuous by their absence, although this, as we have seen, is but a com- paratively recent feature of the region, since most of these forms are represented in the Pliocene of India and Burma. A notable feature of the Oriental as distinct from the Ethiopian region is the circumstance that the great majority of its fruit-bats (like those of Madagascar) belong to the typical genus Pteropus, which is wanting in Ethiopia, while the three genera of the family found in the latter area are absent from the present region. Indeed there is a very curious dissimilarity between the flying-mammals of the two areas, Ethiopia possessing the flying-squirrels of the family Anomaluridce, while the Oriental flying-squirrels all belong to the SriuridiE, the genus Pteromys being peculiar to the region. Among the Insectivora, the so-called flying -lemur (Galeopithecus) is a peculiar Oriental type which has no Ethiopian representative. A somewhat similar instance to that of the flying-squirrels is afforded by the tree-shrews (Tupaiid&) of the Oriental region, which are represented in Ethiopia by the jumping-shrews (Macroscelidida). From the Holarctic region, the Oriental is distinctly differentiated by the presence of apes and lemurs and the abundance of monkeys, together with the presence of the groups mentioned above as being now restricted to this and the Ethiopian region, and likewise by the absence of the typical elaphine deer, marmots, susliks, voles, etc., and the scarcity of sheep and true goats, which, indeed, enter the region only in the north-western frontier of India. Commencing with the man-like apes of the family Simiidce, we find the Oriental region destitute of chimpanzees and gorillas, whose place is taken by the orangs (Simla] of Borneo and Sumatra, characterised by the reddish, instead of blackish, colora- tion of their hair, and their more wide departure from the human type. Orangs appear, however, to have inhabited northern India during the Pliocene, and as chimpanzees were then also in exist- VIII.] PRIMATES. 269 ence there, it would seem that the region must be regarded as the original home of the larger man-like apes. The smaller long- armed apes known as gibbons (Hylobates) are likewise characteristic of the Oriental region, where they range from Assam through the Burmese and Malayan countries to Hainan and Cambodia. In the upper Miocene these apes occurred in Central Europe, but — perhaps on account of their small size — their remains have not hitherto been obtained from the Indian Pliocene. Among the Cercopithecidcz the Oriental genera of monkeys are now entirely distinct from those of Ethiopia, although, as we have seen, the FIG. 60. SLOW LORIS (Nycticebus tardigradus). African genus Papio occurs in the Indian Pliocene and in Madras survived till the Plistocene. And here it may be noticed that this is the only one of the Ethiopian genera of monkeys that is found in Arabia. Of the other genera, Macacus is mainly Oriental, although with representatives in northern Africa, Kashmir, Tibet, and Japan ; while the langurs (Semnopithecus] are likewise almost wholly confined to this region, although they range into Kashmir and eastern Tibet. As both these genera are found in the European and Indian Pliocene, they are evidently ancient types which were formerly widely spread in the eastern half of the Old 2/0 THE ORIENTAL REGION. [CHAP. World. Each represents a sub-family by itself; and as both these sub-families have Ethiopian genera, which are unknown in a fossil state, it may be suggested that the latter (like the wart-hogs) have been evolved within the limits of the Ethiopian region from the Oriental types. To the same sub-family as the langurs belongs the singular proboscis monkey (Nasalis) of Borneo. Among the lemuroid Primates there are Oriental representatives of two families. Of these the Lemurida include the lorises of the genera Nycticebus and Loris, the former ranging over the Burmese, Malayan and Indo-Chinese sub-regions, while the latter is con- fined to southern India and Ceylon. Although these animals are nearly allied to the pottos (Perodicticus) of western Africa, nothing is known as to their past history. The tarsiers (Tarsiidce\ of which there is but a single genus (Tarsius\ although several specific forms have been recognised, are almost confined to the Malayan sub-region, but are represented in Celebes, as they are in the Philippines. Of the Insectivora, the most aberrant and remarkable forms are the flying-lemurs (Galeopithecus), constituting a sub-order by them- selves, and ranging from south Tenasserim through the Malay peninsula and islands to the Philippines. As with the tarsiers, we have no palaeontological history of these creatures, which are probably comparatively modern types. The most characteristic Oriental family of this group is that of the tree-shrews (Tupaiidce], whose members have the form and habits of squirrels, with the structure of shrews. The typical genus Tupaia ' ranges from India throughout the Burmese and Malayan regions, but is unknown in Ceylon ; while the single representative of the pen-tailed tree- shrews (Ptilocercus], characterised by the pen-like extremity of the exceedingly long tail, is confined to Borneo and some of the adja- cent islands. As mentioned in an earlier chapter, the European Miocene genera Lanthanotherium and Galerix appear to be an- cestral types of this family ; and this distribution of the family is a well-marked instance of the curious affinity existing between certain mammalian genera of the middle Tertiaries of Europe and 1 Two species (T. murina and 7". frenata) are often separated as Dendro- gale, but as there is an annectent form (see Thomas, Proc. Zool. Soc. 1892, p. -225), this appears unnecessary. VIII.] INSECTIVORA. 2/1 those of the Malayan sub-region, the absence of Tupaia from Ceylon probably indicating that the genus is essentially a Malayan one which has immigrated but recently into India. The hedgehog family (JErwaceidce), which, as already shown is an ancient one, has a very remarkable distribution in the Oriental region ; true hedgehogs (Erinaceus] ranging into India, but apparently not oc- curring in Ceylon, and being unknown to the west of the Bay of Bengal. In the latter districts their place is taken by the spineless and more rat-like animals forming the genus Gymnura and the FIG. 61. TREE-SHREW {Tupaia tana). allied Hylomys ; and here, again, we have to note the occurrence of an allied type in the European Oligocene, which has been described under the name of Necrogymnurus. In passing, it may be remarked that the survival of these early Tertiary insectivorous types in the Malayan sub-region serves to lend support to the suggestion made in a previous chapter1 that opossums may also have survived in the same area long after they had ceased to exist in western Europe. Passing over the moles (Talpida), of which the typical genus Talpa only just impinges on the region in the frontiers of India> 1 Supra, pp. 51, 57- 2/2 THE ORIENTAL REGION. [CHAP. we come to the Soricidce or shrews. Here the typical shrews (Sorex) are wanting, while the section of the family to which that genus belongs (characterised by the reddish-brown tips to the teeth) is represented only by the genus Soriculus, ranging from the southern slopes of the Himalaya to China. Of the widely-spread musk-shrews (Croddura) it is unnecessary to speak; but it may be mentioned that of the two almost tailless and scaly-footed species forming the genus Annrosorex one is from Assam and the other from eastern Tibet and Pekin. Chimarrogale includes two aquatic shrews, one of which is found in the eastern Himalaya, the hills north of Burma, and M4 Kina Balu in Borneo, while the second is Japanese. Among the Carnivora the region is especially rich in Felidce, containing more species than any other part of the world. The tiger (Felis tigris) is usually regarded as one of the most character- istic mammals of India, but as its range extends northwards to Siberia, while its fossilised remains have been found within the Arctic circle, and it is unknown in Ceylon, there is a great proba- bility that this feline is a comparatively recent immigrant into India from the north-east. The range of the lion (F. leo) in this •region is limited to India, not extending to the eastward of the Bay of Bengal, and as this animal was widely distributed during the Plistocene in Europe, while it ranges all over Africa, it may be regarded as essentially a western type, or exactly the opposite of the tiger. Possibly certain remains from the Indian Plistocene which have been assigned to the latter animal may really belong to the former. As noticed on p. 234, there are other species of Felis, as well as the hunting-leopard (CynteJurus), which are com- mon to India and Africa, some of these occurring in the European Plistocene, while only the jungle-cat (F. chaus) is found to the eastward of the Bay of Bengal, and there not further east than Burma. On the other hand, there are certain species, like the clouded leopard (F. nebulosa) and the marbled cat (F. marmorata), which are essentially eastern forms, their range including the Malayan sub-region and India, but not Ceylon. The rusty- •spotted cat (F. rubiginosa) and the Indian desert-cat (F. ornatd] are species whose range is limited in one case to India and •Ceylon, and in the other to India alone. VIII.] CARNIVORA. 2/3 In the civets and their allies ( Viverridtz] the Oriental region approaches the Ethiopian in richness, and thereby stands in marked contrast to the Holarctic, which contains only a single species of Genetta and another of Herpestes in southern Europe, although the latter genus ranges into Kashmir. Of the true civets ( Viverrd] the whole of the species, with the exception of one from the Ethiopian region, are Oriental, and some are confined to the countries to the east of the Bay of Bengal; one small species being separated by many zoologists as Viverricula. The beautifully- coloured linsangs (Linsangd) are exclusively Oriental, and are con- fined to the eastern Himalayan and Malayan sub-regions, although represented in West Africa by the nearly allied Poiana. Equally characteristic of the region are the two species of Hemigale, which are, however, exclusively Malayan, H. hosei being limited to the mountains of North Borneo. The palm-civets of the genus Para- doxurus range throughout the region, and have also representatives in Celebes : their place being taken in the Ethiopian region by the allied genus Nandinia. On the other hand, the two species of small-toothed palm-civets constituting the genus Arctogale are restricted to the Burmese and Malayan sub-regions ; the same being also the case with the binturong, which is the sole repre- sentative of the genus Arctitis, distinguished by its pencilled ears and prehensile tail. Still more circumscribed is the range of the peculiar genus Cynogale, of which the single species is confined to the Malayan sub-region. All the foregoing forms belong to the sub-family Viverrina \ the Herpestincz, which are so numerous in the Ethiopian region, being represented only by species of the large and widely-spread genus Herpestes. Both this genus and Viverra date from the European Oligocene, the latter also occurring in the Pliocene of France and India ; but none of the others are known in a fossil state. It is, however, probable that most of the other genera are comparatively modern derivatives from the original stock ; and the high development in the Malayan sub-region of a group first known from the Oligocene of Europe is another instance of the relationship of the faunas of these countries. Although the striped hyaena (Hycena striata) is by no means confined to India, its range extending through south-western Asia L. 18 274 THE ORIENTAL REGION. [CHAP. to northern Africa, it is unknown in Ceylon or in the countries on the eastern side of the Bay of Bengal, which forms, indeed, the present limits of the range of the genus in this direction. As remains of the existing African spotted hyaena (H. crocutd) have been met with in a cave in Madras, while they are common in the Plistocene of southern and central Europe, it is manifest that both these animals are as essentially western types as is the lion. And it is a curious circumstance that nearly all these western types of mammals ranging into India (of which a list is given in the sequel) belong to genera which date only from the Miocene or Pliocene, whereas very many of the Malayan or eastern types date from the Oligocene. During the Pliocene a single species of hyaena ranged as far eastwards as China, and species were exceedingly abundant in India at the same epoch. As regards its Canidcz, the Oriental region is inferior to the Ethiopian in lacking any peculiar generic type, although it possesses a true wolf (Cants pallipes), and three species of wild dog (C. rutilans, etc.), the latter, on account of the absence of the last tooth in the lower jaw and other differences, being frequently referred to a distinct genus, under the title of Cyan. Whereas, however, the Indian wolf, which ranges into southern Arabia, is unknown either in Ceylon or in the countries to the east of the Bay of Bengal, the wild dogs are found throughout the region, and have also a representative beyond it in the mountains of Central Asia, and they are likewise known by fossil species from the European Plistocene. The wolf, which is very closely allied to the European species, may be the descendant of a fossil form found in the Siwaliks, but its absence from Ceylon would seem to indicate that it has only reached southern India at a comparatively modern date. No foxes are known to the east of the Bay of Bengal, and the jackal does not range east of Burma. The Oriental region is the home of three well-marked species of bears, and thereby presents a decided contrast to the Ethiopian. Of these the Himalayan black bear (Ursus torquatus) ranges from the forest districts of the Himalaya to Burma, and thence to the Indo-Chinese sub-region. The small Malayan bear (U. malay- anus] is restricted to the Burmese and Malayan sub-regions ; and the great Indian sloth bear (Melursus ursinus], which is the sole VIII.] CARNIVORA. 275 representative of a separate genus, is confined to India and Ceylon, and is known to have been an inhabitant of Madras since the Plistocene era. Not improbably it may be the descendant of a Siwalik species ( U. theobaldi), which is the earliest known repre- sentative of the true bears; and the Malayan species may be FIG. 62. INDIAN SLOTH-BEAR (Melursus ursinus). derived from a small extinct bear whose remains occur in the Plistocene of the Narbada valley. One of the most remarkable of Oriental carnivores is the panda, or cat-bear (sElurus fulgens), which ranges from the East- ern Himalaya to Yunnan, and is the single existing Old World representative of the Procyonida. Curiously enough, the remains of a much larger species of the same genus have been discovered in the English Pliocene ; and it is thus evident that sElurus formerly enjoyed a wide range. From the restriction of all the other known members of the family to the New World, fossil types may be looked for in eastern Asia, as it is quite clear that the distributional area of the group must once have been con- tinuous. In the Mustelidce there are four generic types very character- 18— 2 276 THE ORIENTAL REGION. [CHAP. istic of the region, although two of them are not confined to it. The first of these comprises the three species of sand-badger (Arctonyx), two of which are found in the Himalayan and Burmese sub-regions, while the third is Tibetan. The single Oriental species of ratel (Mellivora) is restricted to India, exclusive of Ceylon ; a fossil species occurring in the Siwaliks. The only other living form is Ethiopian. Its distribution would thus seem to indicate that the genus originated in northern India, whence it FIG. 63. INDIAN RATEL (Mellivora ratel). migrated into Africa while there was a free communication between the two continents, and that it only reached southern India (where it is unknown on the Malabar coast) at a comparatively recent epoch. The third genus, Helictis, comprising badger-like animals with long bushy tails, is represented by four species, ranging from India to China, but unknown in Ceylon. Lastly, the teledu, or small burrowing badger (Mydaus meliceps) of the Malayan sub- region, is the sole representative of its genus, and is found at con- VIII.] RODENTIA. 277 siderable elevations in Java, as well as in Sumatra and Borneo. No fossil representatives of either of these two genera are at present known. Among the rodents, the grooved-toothed squirrel (Rhithro- scturus) is a peculiar type confined to the island of Borneo; and the pigmy squirrels (Nannosciurus) are represented by four Malayan species, the only other member of the genus being West African. The true squirrels (Sciurus), as mentioned above, attain their maximum development in the Malayan sub-region. The flying - squirrels are represented by the genera Pteromys and Sciuropterus, the former being exclusively Oriental, and the latter having one species in the eastern, and a second in the western half of the Holarctic region, in addition to being represented in a fossil state in the French Miocene. In the Muridcz there are no less than eleven genera — in most cases respectively represented by only a single species — peculiar to this region, while another is Oriental and Ethiopian only. Of the peculiar types, one of the most remarkable is Chrotomys, from the mountains of Luzon, in the Philippines, belonging to the sub-family Hydromyince, of which the typical forms are Australian1. From other members of the sub-family the single species of this genus differs in having three (in place of only two) pairs of molar teeth, thereby forming a link with ordinary murines. This animal, which is about the size of a rat, is easily recognised by the presence of an orange or buff line running down the middle of the back. Luzon has also yielded another rat, provisionally referred to the Australian genus Xeromys*. Another unique Oriental type is found in the long- tailed dormouse-like form from the Malabar coast known as Platacanthomys, which constitutes a sub-family by itself. Phlao- mys, likewise representing a separate group of the same rank, is restricted to the Philippines, and is characterised by the molar teeth being divided into three transverse lobes. The one species is of very large size. Nearly allied is a huge, rough-haired, grey or blackish rat, from the mountains of Luzon, which may be compared in size to a small marmot, and for which the name Crateromys has been suggested. This differs from Phlceomys by 1 Vide suprh, p. 40. 2 Appendix, No. 31 2/8 THE ORIENTAL REGION. [CHAP. the smaller claws and more bushy tail, and also by the completely tuberculate molars. The single species of the Burmese rat-like Hapalomys differs from all other members of the sub-family Murincz in possessing three longitudinal rows of tubercles on the lower as well as on the upper molar teeth. The one repre- sentative of the allied genus Vandeleuria ranges from India and Ceylon to Yunnan. The pencil-tailed tree-mice (Chiropodomys), of which there are three species, are restricted to the Burmese and Malayan sub-regions; and the small red rat representing the genus Pithechirus is known only from Sumatra and Java. With the shrew-rat (Rhynchomys] we revert to several peculiar forms from the mountains of Luzon, in the Philippines. This rodent, which is about the size of an ordinary rat, has the muzzle extraordinarily slender and elongated, with very feeble incisors, and it is pro- bable that it lives on animal substances, possibly caterpillars. The two other Philippine types form the genera Carpomys and Batomys ; the former with two, and the latter with a single species. They are more or less dormouse-like forms, with blunt muzzles, thick woolly fur. and long and well-haired tails. Lastly, the bush-rats (Golunda) have one Indian and another Ethiopian representative. An interesting instance of how the present distribution of a genus is explained by palaeontological discoveries is afforded by the brush-tailed porcupines (Atherura], now represented by one species from the Malayan, and a second from the West African sub-region ; the connecting form being one of which fossil teeth have been found in the Karnul district of Madras. From this it may be inferred that the genus was probably also represented in the Siwaliks. To the same family (Hystricida) belongs a peculiar porcupine from Borneo, constituting the genus Trichys. Passing on to the ungulates, we have first to notice a peculiar group of oxen forming a section of the genus Bos, which is con- fined to this region, and characterised, in addition to certain features of the skull and horns, by the dark colour of the males, or of both males and females. Of these, the gaur (B. gaurus) inhabits both India and the Malayan countries, but appears never to have reached Ceylon ; while the banteng (B. sondaicus) is confined to the countries on the east of the Bay of Bengal. Fossil representa- VIIL] UNGULATA. 2/9 tives of this group occur in the Indian Plistocene ; and certain generalised oxen from the Siwalik Hills and the Pliocene of southern Europe, in which the females were generally or always hornless, may have been the ancestral type. The Indian buffalo (B. bubalus] is markedly distinct from the Ethiopian forms, and has ancestral representatives in the Indian Pliocene and Plisto- cene. While abundant in Ceylon, it is probably unknown in a truly wild state to the east of the Bay of Bengal. The Philippine buffalo, or tamarao (B. mindorensis) is regarded by some as a FIG. 64. JAPANESE SEROW (Nemorhadiis crispus}. cross between the last and the anoa of Celebes ; ancestral types of the latter occurring, as already mentioned, in the Siwaliks. In the same family the short-horned goats of the genus Hemitragus are represented by two Indian species, one inhabiting the Himalaya and the other the Nilgiris ; the third living species being south Arabian. One extinct species occurs in the Siwaliks and a second one in Perim Island, so that the group is essentially an Indian one ; and, as already mentioned, its present distribution can only 280 THE ORIENTAL REGION. [CHAP. be accounted for by a lowering of the temperature during a past epoch. Of the goat-like genera Nemorhcedtis and Cemas, the former has a wide range in the region and also extends into northern China and Japan, while the latter is represented solely by the Himalayan goral ; no fossil types of either being known. In its poverty of antelopes (exclusive of the widely-spread gazelles) the Oriental presents a most remarkable contrast to the Ethiopian region, although this poverty is largely a feature of the present epoch, African types being common in the Siwaliks. The sole existing forms are the four-horned antelope (Tetraceros quadri- cornis], the black-buck (Antilope cervicapra)^ and the nilgai (Bos- elaphus tragocamelus], each of which forms a genus by itself, and all of which are restricted to India, exclusive of Ceylon. Indeed, it is a remarkable feature that true antelopes and gazelles are unknown to the eastward of the Bay of Bengal ; although this may be chiefly or entirely due to the countries to the eastward being unsuited to their habits. The nilgai, which has fossil representa- tives in the Indian Plistocene and Pliocene, is allied to the kudu group of Africa, while the four-horned antelope is a near relative of the duikers. It will be unnecessary to say anything with regard to the true goats (Caprd) and sheep (Ovis) inhabiting the region, since these are found only on the north-western frontier of India, and are obviously intruders from the Holarctic region. It is, however, important to mention that extinct representatives of one, if not of both groups, occur in the Siwalik Hills. The abundance of Cervidce is one of the most noticeable features distinguishing the Oriental from the Ethiopian region ; there being an equally marked difference in this respect between the former and the Holarctic area. Although the majority of the Oriental deer are now included in the genus Cervus, the typical, or elaphine group, as represented by the red deer and the wapiti, is entirely wanting, its place being taken by the sambur and its allies (C. unicolor), forming the rusine group; the swamp-deer (C. duvaucelt), which with another species constitutes the rucervine group, and the Indian spotted deer (C. axis), alone representing the axine group. Rusine deer are abundant in the Indian Siwaliks, but appear to be unknown in the Pikermi beds. Although they have one Tibetan representative, the smaller deer VIII.] UNGULATA. 28 1 known as muntjacs (Cervulus) — which are characterised by the length of the pedicles of the antlers and the shortness of the antlers themselves — form a very characteristic Oriental group, ranging over the entire region. Not improbably they are represented in the Pliocene of Europe. The chevrotains, or TragulidcK^ which have already been shown to be abundant in the European Oligocene and Miocene — remains of the West African genus occurring in the latter deposits and the Indian Pliocene — are represented in the Oriental region by Tragulus, which dates from the Siwalik epoch, and ranges from India and Ceylon to the Philippines. Although wild camels are now everywhere unknown, it is probable that India and the Holarctic region was their original home, remains of the genus Camelus being found in the Pliocene of the Siwalik Hills. The large number of species of true pigs (Sus} characterising the Oriental region is a notable feature, India itself being in- habited by a species (Sus cristatus] nearly allied to the European wild boar, while the Malayan sub-region is the home of a consider- able number of species differing more or less markedly from the latter. The genus is well represented both in the Pliocene and Plistocene of India, but in neither of these formations are there any of the Ethiopian types of the family. With the exception of the Ethiopian, the Oriental region is now the sole one where the family Rhino cerotida still exists ; but there is a remarkable difference between the species inhabiting the two areas, all the three living Asiatic forms being furnished with teeth in the front of the jaws, which, as we have seen in the last chapter, are wanting in the African species. While one of the Oriental rhinoceroses (R. sondaicus} ranges from eastern Bengal to the Malayan islands, and a second (R. sumatrensis) from Assam to the same, the great Indian species (R. unicornis) is unknown to the eastward of Assam, as it is in Ceylon. Fossil remains of the latter are found in the Plistocene of the Narbada valley, while ancestral types both of this species and of R. sondaicus are met with in the Pliocene of the Siwalik Hills. It is, however, very remarkable that Ethiopian types of the genus occur not only in the last-named deposits, but likewise in the Plistocene of Madras ; the total extinction of this group in India being, as in the case of 282 THE ORIENTAL REGION. [CHAP. other Ethiopian types, almost impossible to account for. One of the two-horned extinct Indian rhinoceroses (R. platyrhinus] ap- pears to have been the ancestor both of the existing R. simus of Africa and R. antiquitatis of the Plistocene of northern Asia and Europe ; the evolution of the latter species having not improbably taken place in the countries lying between India and China, whence the creature wandered northwards and westwards with the mammoth to the Arctic tundras. With regard to the Equidce, it will suffice to mention that species of Equus occur in the Plisto- cene of Central India and Madras, and that wild asses (of a type markedly different from the African wild ass) occur in Sind and Kach. The genus, like the antelopes, is, however, totally un- known in the countries to the east of the Bay of Bengal, as it is in Ceylon. In the Tapiridcz, the Malayan tapir (Tapirus indicus) inhabits the Malay Peninsula as far north as Mergui, and also the islands of Sumatra and Borneo. It is important to notice that although fossil remains of tapirs are unknown from the Pliocene of the Siwaliks, they are met with in caverns in China. Distinguished, among other features, from its African cousin by the thinner and more numerous enamel-plates of its molar teeth, the Indian elephant (Elephas indicus) ranges over the greater part of the region, being found in suitable districts in India, Ceylon, Burma, the Malay Peninsula, Cochin China, and Sumatra. This species is a near ally of the mammoth (E. primigenius) ; and it may prove that both are descendants of a Siwalik species (E. hysudricus], which has molar teeth of the type we should expect to find in such an ancestral form. If this view be correct, the mammoth has probably wandered to northern Europe and Siberia from the countries lying just to the east of India. It has been mentioned in an earlier chapter that the extinct so-called stego- dont elephants (such as E. clifti and E. insignis] are mainly confined to this region, although some of the species are found in north China and Japan. As these elephants form the transition between Elephas and Mastodon, and also since the species of the latter genus which may be regarded as the original stock of the elephants is confined to the Indian and Malayan Pliocene, it may be taken for granted that the elephants have been developed from the mastodons within the limits of the Oriental region. In the VIII.] PANGOLINS. 283 Plistocene of the Narbada Valley in central India there occurs a species (E. namadicus] closely allied to the contemporary Euro- pean E. antiqims, in both of which the molars are intermediate in structure between those of the living Indian and African species. Elephas planifrons of the Siwaliks, which has molars of a still more generalised type, is equally closely allied to E. meridionalis of the upper Pliocene of Europe; and it is quite probable that the former may be the original ancestral stock of the African elephant. It is worth mentioning that the stegodont elephants survived till the Plistocene ; and also that some of the species of this group inhabiting India, as well as certain mastodons, ranged as far east- wards as Java, Borneo, China, and Japan. FlG. 65. WHITE-BELLIED PANGOLIN (Manis tricuspis). The last mammals that we have to mention are the pangolins (Mamdcz), which are now common to the Oriental and Ethiopian regions, and appear to be represented by an extinct genus in the European upper Oligocene. The presence of horny scales in- vesting the whole of the body and tail serves to distinguish the 284 THE ORIENTAL REGION. [CHAP. pangolins from all other mammals whatsoever ; and the Oriental species are further characterised by having the median series of scales on the body continued to the tip of the tail, and likewise by the presence of numerous isolated hairs between the scales of the back, as well as by the presence of small ears. Of the three Oriental species, Manis javanica ranges from Burma through the Malay Peninsula to Java and Borneo ; M. aurita extends from Nipal to the Indo-Chinese sub-region ; while M. pentedactyla is restricted to India and Ceylon. The most remarkable feature connected with the distribution of the group is, however, the circumstance that claw-bones indistinguishable from those of the giant pangolin (M. giganted] of West Africa have been discovered in a cavern in the Karnul district of Madras. In the following list the leading results of the foregoing survey are put in tabular form, the italics indicating groups or species peculiar to the region. I. Primates. Simia. Borneo and Sumatra ; fossil in India. Hylobates. Burmese and Malayan ; fossil in European Miocene. CERCOPITHECID^E. Macacus. Now mainly Oriental, but occurring on the southern borders of the Holarctic region ; fossil in the European and Indian Pliocene. Semnopithecus. An outlying species in Eastern Tibet and one in Kashmir ; fossil in Pliocene of Europe and India. Nasalis. Borneo. Nycticebus. Burmese, Malayan, and Indo-Chinese. Loris. S. India and Ceylon. Elsewhere only in Celebes. Tarsius. Malayan, extending into Celebes. VIII.] LIST OF THE FAUNA. 285 II. Insectivora. GALEOPITHECID&. Galeopithecus. Malayan. TUPAIIDsE. Tupaia. Indian and Malayan. Ptilocercus. Borneo and some adjacent islands. ERINACEID^E. Gymnura. Burmese and Malayan. Hylomys. Burmese and Malayan. SORICID/E. Soriculus. Himalayan and Indo-Chinese. Anurosorex. Known by one species from Assam and a second from Tibet and Pekin. Chimarrogale. Represented by one species from the eastern Himalaya, hills north of Burma, and Mt Kina Balu, Borneo, and a second from Japan. III. Carnivora. FELID^:. Very numerous in the region. Cynaelurus. Indian and Ethiopian ; the one species being common to the two areas ; fossil in Indian Pliocene. VlVERRID/E. Viverra. All the species, except a single Ethiopian one, are Oriental, one of these being frequently regarded as the representative of a distinct genus ( Viverricula) ; fossil in European Oligocene and European and Indian Pliocene. Hemigale. Malayan. Linsanga. Malayan and E. Himalayan. Paradoxurus. An outlying species in Celebes. Arctogale. Burmese and Malayan. Arctictis. Burmese and Malayan. Cynogale. Malayan. URSID^E. Melursus. India and Ceylon. 286 THE ORIENTAL REGION. [CHAP. III. Carnivora. (Cont.) PROCYONID^E. sElurus. Eastern Himalayan and Burmese ; fossil in English Pliocene. MUSTELID^E. Arctonyx. Two E. Himalayan and Burmese species, and probably a third from Tibet. Mellivora. One Indian and another Ethiopian species ; fossil in Indian Pliocene. Helictis. India to China. Mydaus. Malayan. IV. Rodentia. SCIURID^E. Rhithrosciurus. Borneo. Nannosciurus. Represented elsewhere by a single West African species. Sciurus. This almost cosmopolitan genus appears to attain its maximum development in the Malayan sub-region. Pteromys. MURID^E. Chrotomys. Philippines (Luzon). Xeromys. Philippines (Luzon), and Australia. Phlceomys. Philippines. Crateromys. Philippines (Luzon). Hapalomys. Burma. Vandeleuria. India and Burma. Chiropodomys. Burmese and Malayan. Pithechirus. Sumatra and Java. Rhynchomys. \ Carpomys. >• Philippines (Luzon). Batomys. / Golunda. One Indian and one Ethiopian species. HYSTRICID^E. Atherura. One Malayan and one West African species ; fossil in Indian Plistocene. Trichys. Borneo. VIII.] LIST OF THE FAUNA. 287 V. Ungulata. Bos. The Bibovine group exclusively Oriental. Hemitragus. Two Indian species, and a third in the South Arabian sub-region of Ethiopia; fossil in Indian Pliocene. Nemorhaedus. Largely Oriental (Himalayan, Bur- mese, and Malayan), but extending into northern China and Japan. Cemas. Himalayan. Tetraceros. Indian. Antilope. Indian ; fossil in Plistocene. Boselaphus. Indian ; fossil in Plistocene and Pliocene. Cervus. The Rusine, Ruce.rvine, and Axine groups of this genus are characteristic of the" region, although the first is also represented in the Austro- Malayan. Cervulus. Mainly Oriental, but with one Tibetan species. TRAGULID^E. Tragulus. India, Ceylon, and Malayan sub-region ; fossil in Indian Pliocene. Sus. Attains its maximum specific development in the Malayan sub-region. RHINOCEROTID^E. Rhinoceros. The three existing Oriental species differ from the Ethiopian forms in having front teeth. TAPIRID^E. Tapirus. Malayan : elsewhere living only in the Neogaeic realm, but widely distributed in a fossil state, although absent from the Siwaliks. 288 THE ORIENTAL REGION. [CHAP. V. Ungulata (cont.). ELEPHANTID^E. Elephas. The existing Oriental elephant is widely different from the Ethiopian, although nearly allied to the Holarctic mammoth ; the extinct Stegodont group is mainly Oriental, although extending into north China and Japan. VI. Effbdientia. MANID^E. Manis. Elsewhere only in Ethiopian region ; fossil in Indian Plistocene. The relations of peninsular India to the Himalayan area have been already discussed at the commencement of of^hV Regiony this cnaPter \ wm'le tne land-connection which appears to have existed between India and Mada- gascar, and thus with Africa, has been alluded to in an earlier one. The latter connection must have ceased to exist before the Pliocene era ; and, as we have seen, the descendants of the Siwalik mammals would appear to have made their way into Ethiopia across Syria or Arabia. During the Pliocene, India, at least, could not have been distinguished as a region from Ethiopia as it exists at the present day ; and even in the Plistocene the connection between the faunas of the two areas was much more intimate than it is now. The full reason for this gradual dis- appearance of the modern Ethiopian types from the Indian area will probably never be known ; but there can be little doubt that the gradual refrigeration of the northern hemisphere with the advent of the glacial period has been largely instrumental ; the present distribution of Hemitragus being only explicable on the hypothesis of a marked lowering of the temperature over India. The more peculiar mammals now inhabiting the Oriental region may be roughly arranged under five headings. The first will include those that are common to India and some of the countries to the west or south-west, but are, for the most part, unknown in either Ceylon or the countries to the eastward of the Bay of Bengal. Under this category may be included the follow- ing, viz. :— VIII.] LIST OF THE FAUNA. 289 INSECTIVORA. Erinaceus. CARNIVORA. Felis leo. Ethiopian, and European Plisto- cene. Felis chaus. Ethiopian, and European Plistocene ; ranges eastward into Burma. Felis caracal. Ethiopian. Cynaelurus jubatus. Ethiopian. Hyaena striata. Western Asia and North Africa. Hyaena crocuta. Ethiopian ; occurs both in India and Europe during the Plistocene. No representative of the genus known in the Burmese or Malayan countries. Canis aureus. S. W. Asia ; also ranges into Burma. Canis pallipes. South Arabian. Mellivora. The Indian and Ethiopian species very closely allied. RODENTIA. Golunda. Ethiopian ; the Indian species ranges into Ceylon. UNGULATA. Hemitragus. South Arabian. Antelopes. ) None known east of the Bay of Equus. j Bengal. The second group includes such genera and species as are common to India and Ceylon, but are unknown elsewhere. Here we have : — PRIMATES. CARNIVORA. RODENTIA. UNGULATA. Loris. Felis rubiginosa. Melursus. Golunda ellioti. Bos bubalus. Cervus axis. Tragulus memimna. EFFODIENTIA. Manis pentedactyla. L. 290 THE ORIENTAL REGION. [CHAP. The third group, which is a small one, comprises types which are confined to India ; it includes Felis ornata. Boselaphus. Tetraceros. Antilope. Rhinoceros unicornis. In the fourth group we have generic or specific forms common to India and the countries to the eastward of the Bay of Bengal, but unknown in Ceylon or in the countries to the west or south- west. This list comprises the following, viz. : — PRIMATES. Hylobates. INSECTIVORA. Tupaia. CARNIVORA. Felis tigris. Felis nebulosa. Felis marmorata. ^Elurus. Helictis. Arctonyx. UNGULATA. Bos gaums. Bos frontalis. Nemorhaedus. Cervus porcinus. Finally, we have (among others) the following group of genera and species confined to the countries lying immediately to the eastward of the Bay of Bengal, viz. :— PRIMATES. Simia. Nasalis. Nycticebus. Tarsius. INSECTIVORA. Galeopithecus. Gymnura. Hylomys. CARNIVORA. Mydaus. RODENTIA. Rhithrosciurus. Trichys. VIII.] AFFINITIES OF THE FAUNA. 2QI UNGULATA. Bos sondaicus. Tragulus javanicus. Tapirus. EFFODIENTIA. Manis javanica. Other forms might be added to several of these lists, but those included are sufficient for the purpose of showing that the present mammalian fauna of India is a complex formed by an admixture of western and eastern types. The first group is an essentially modern one, all the generic types contained in it, with the exception of Erinaceus (which dates from the Miocene), being unknown before the lower Pliocene ; while, if we except Erinaceus and Golunda, all occur in the Siwaliks. In the Carnivora, there is evidence of all the species except the first three being descended from Siwalik ancestors ; and it is quite probable that the three species of Felis may trace their origin to felines which lived either in the Siwaliks or Persia during the Pliocene, in which event the lion, and not the tiger, should be regarded as the characteristic large Indian feline. With the probable exception of Loris, the second group is also a modern one ; all the forms save Loris and Golunda having ancestral types in either the Pliocene or Plistocene of India, and none of the genera being known before the former epoch. And it may be mentioned here that the absence of so many of the smaller types of Oriental mammals from the Siwaliks is no indication that the genera did not flourish in India during the Pliocene age, but is due to the strata being unsuited to the preservation of their remains. The remarks applicable to the second group will like- wise befit the third. On the other hand, the fourth and fifth groups appear to have less connection with the Siwaliks, while several of the types are older ones. For instance, we have no proof of the existence of oxen nearly allied to Bos gaurus in the Siwaliks, although such are found in the Indian Plistocene ; neither is there any evidence of a Siwalik tapir. Tupaia is a near relative of the European Miocene Galerix and the Oligocene Lanthanotherium ; as is Gymnura of the Oligocene Necrogymnura ; while Hylobates is 19—2 THE ORIENTAL REGION. [CHAP. represented in the European Miocene1. The reasons for regard- ing the tiger as a comparatively modern immigrant into southern India have already been stated. A Siwalik origin is, however, indicated for Simla ; but concerning the other genera the palaeon- tological history is unfortunately a blank. The affinity between the faunas of West Africa and the Malayan sub-region has been already alluded to ; but there are also indica- tions of a connection between that of the latter area on the one hand and that of southern India and Ceylon on the other, as exemplified by the occurrence of Nycticebus in the Malayan sub- region and of Loris in south India and Ceylon. To explain the latter connection, Dr Blanford2 has discussed the possible exist- ence of a direct land-communication between the two areas ; but this connection, as he admits, scarcely seems necessary, since in such cases the true explanation would seem to be the survival of old types in the tropical forest-regions. And here it may be noticed that the Malayan types common to, or represented by allied forms in West Africa, are such as either have representatives in the Indian Pliocene or Plistocene, or such as we might naturally expect to meet with there if small forms were commonly preserved. For instance, Simia and Anthropopithecus, and Dorcatherium and Tragulus, are all represented in the Siwaliks, and Atherura occurs in the Madras Plistocene. This being so, what is more likely than that lorises, linsangs, and palm-civets, of types more or less intermediate between the existing Malayan and West African representatives of those groups, should have flourished in India during the Pliocene era ? Nannostiurus, again, should certainly be expected to occur in the Indian Pliocene. On this point I think Dr Wallace3 was on the right track when, in writing of the Malayan sub -region, he observed that "Here alone, in the Oriental region, are found the most typical equatorial forms of life-organisms adapted to a climate characterised by uniform but not excessive heat, abundant moisture, and no marked departure from the average meteorological state throughout the year. These 1 If my memory serves me right, I have been shown teeth from the upper Oligocene Phosphorites of France closely resembling those of Hylobates. 2 Manual of Geology of India, rst Ed. pt. 1. p. Ixviii. 3 Geographical Distribution of Animals, vol. I. p. 335. VIII.] AFFINITIES OF THE FAUNA. 293 favourable conditions of life only occur in three widely separated districts of the globe — the Malay Archipelago, Western Africa, and equatorial South America. Hence, perhaps, it is that the tapir and trogons of Malacca should so closely resemble those of South America ; and that the great anthropoid apes and crested horn- bills of Western Africa should find their nearest allies in Borneo and Sumatra." In addition to the resemblances between the mammalian fauna of the Malayan sub-region and that of West Africa, there are, however, equally well-marked differences between the former and that of Ethiopia in general. Among Burmese and Malayan types wanting in Africa, we have especially to note Tapirus, Gymnura, Tupaia, Hylobates, and sElurus. From the small size of their representatives it would be unfair to argue anything from the absence of the last four of these from the Siwaliks, but the case is very different with regard to Tapirus, which ought surely to have been found did it exist there. As this genus is equally wanting in the Pikermi and Persian Pliocene, while it occurs in that of France, Germany, England and China, it is a fair inference that it has reached the Malayan countries by a route lying north of India. And the occurrence of sElurus in the English Pliocene suggests that the same may be the case with that genus. If this be so, it is not an improbable hypothesis that the other genera men- tioned, all of which have representative types in the European Tertiaries, may have migrated eastwards by a similar route ; and in this connection it is especially noteworthy that such of the genera in question as enter India at all, occur only in the eastern or southern districts. With regard to the date of the separation of Ceylon from India, the numerous species of mammals common to the two areas show that this must have taken place at a very recent date, comparatively speaking ; although, as aforesaid, at a period when several of the mammals now inhabiting southern India had not yet occupied that portion of their distributional area. When discussing the possibility of a former land-connection across the Bay of Bengal between Ceylon and southern India on the one hand, and the Malayan countries on the other, Dr Blanford was careful to point out that the ocean-bed afforded no evidence 294 THE ORIENTAL REGION. [CHAP. in favour of such a line of communication. This feature, together with certain marked differences between the mammals of the two areas, appears to afford a conclusive argument that these countries have never been much more closely connected than they are at present. Had any more extensive connection existed, we should surely expect to find antelopes, gazelles, and perhaps asses, in the more open districts of upper Burma ; while the Bay of Bengal would scarcely have formed such a sharp line limiting the eastward range of wolves, foxes, hyaenas, and the other mammals mentioned in the list on page 289, as it actually does. That list is confined to existing types, but if fossil forms had been included, Hippo- potamus might have been added, since the extinct Oriental representatives of that genus do not range further east than Burma (whither they evidently migrated down the river-valleys from northern India), no species being known from the Tertiaries of China, Japan, or the Malayan islands. These circumstances, together with the depth of the sea in the Bay of Bengal, seem to disprove the suggestion of Dr Wallace1 that a continuous tract of land formerly connected Borneo and the rest of Malaysia with the central peak of Ceylon, and extended eastwards to Hainan. Within the limits of the present volume it would be quite impossible to give a detailed description of the SuWegion mammalian faunas of the Malay Peninsula and Islands ; and I have accordingly selected that of the Bornean group, as an example of what may be called the typical Malayan sub-region, as distinct from Java, which differs markedly in its fauna from Borneo and Sumatra. The chief reason for selecting Borneo is that its fauna has been carefully worked out by Mr A. H. Everett2 and Mr C. Hose3, from whose papers the following list of mammals (exclusive of bats), with some emendations and additions, has been compiled ; species, like the rat, mouse, and buffalo, which have obviously been introduced, being omitted. Mr Everett includes within the Bornean group the island of Palawan, and states that the group may be defined " by a line which starts from a point immediately to the west of 1 Op. dt. p. 359. 2 Appendix, No. 14. 3 Descriptive Account of the Mammals of Borneo, Diss, Norfolk, 1893. VIII.] SUB-REGIONS. 295 St Julian Island in the Tambelan Archipelago, and being drawn south of the Great Natuna (Bungoran Island), passes northward of Labuan and thence follows the loo-fathom line, so as to embrace Balabac, Palawan (Paragua), the Calamianes, and the Cuyo Islands, and, returning along the same line of soundings on the southern side of Palawan, is drawn immediately to the islands of Cagayan Sulu, and Sibutu, — whence it is continued through the Makassar Straits south of the Paternoster, Lauriot (Laset Ketjil), and Solombo islets, and in a north-westerly direction through the Karimata Strait back to the island of St Julian." In the following list the genera and species peculiar to the group are printed in italics, those which are confined to the Palawan sub-group having that name placed after them in brackets. The distribution of the other forms has been indicated as far as practicable. PRIMATES. Simia satyrus. Sumatra. Hylobates leuciscus. Java to Philippines. ,, muelleri. Semnopithecus maurus. Malay Peninsula and Java. ,, chrysomelas. crudger. „ hosei. ,, everetti. ,, rubicundus. ,, frontatus. Nasalis larvatus. Macacus arctoides. Burma, China, E. Tibet, etc. „ nemestrinus. Burma, Malay Peninsula, Sumatra, and Java. „ cynomolgus. Burma to Philippines. Nycticebus tardigradus. Burma to Philippines. Tarsius spectrum. Sumatra, Java, and Banka. INSECTIVORA. Chimarrogale himalayica. E. Himalaya and hills north of Burma. Crocidura fuliginosa. E. Himalaya. „ faitida. „ dor ice. „ indica. hosei. 296 THE ORIENTAL REGION. [CHAP. INSECTIVORA (cont.). Gymnura rafflesi1. S. Tenasserim, Malay Penin- sula, and Sumatra. Hylomys suilla. Burma, Malay Peninsula, and Sumatra. Ptilocercus lowi. Also in some neighbouring islands. Tupaia murina. ,, javanica. Java and Malay Peninsula. „ minor. „ gracilis. ,, melanura. ,, ferruginea. Burma, Malay Peninsula, Suma- tra, and Java. ,, splendidula. „ tana. Sumatra, Natuna Islands. ,, dor satis. „ picta. ,, montana. Galeopithecus volans. Malay Peninsula, S. Tenas- serim, Siam, Sumatra, and Java. CARNIVORA. Felis planiceps. Malay Peninsula and Sumatra. ,, badia. „ temmincki. E. Himalaya, Burma, Malay Peninsula, and (?) Sumatra. ,, bengalensis. India to Philippines. ,, marmorata. E. Himalaya to Sumatra. ,, nebulosa. E. Himalaya to Formosa. Viverra tangalunga. Malay Peninsula, Sumatra, Philippines, and Celebes. Linsanga gracilis. Java and (?) Sumatra. Paradoxurus leucomystax. Malay Peninsula and Sumatra. ,, hermaphroditus. India to Java and Sumatra. ,, philippinensis. Philippines. Arctogale leucotis. Sikhim, Burma, Malay Penin- sula, and Sumatra. 1 Mr Jentink regards the Bornean form as a distinct species (G. alba]. VIII.] SUB-REGIONS. 2Q/ CARNIVORA (cent.}. Hemigale hardwickei. Malay Peninsula and Su- matra. ,, hosei. Arctictis binturong. E. Himalaya, Burma, Siam, Malay Peninsula, Sumatra, and Java. Helictis everetti. Cynogale bennetti. Malay Peninsula and Su- matra. Herpestes brachyurus. Malay Peninsula. ,, semitorquatus. Sumatra1. (?) Canis rutilans. Malay Peninsula, Sumatra, and Java. Ursus malayanus. Arakan, Tenasserim, Malay Peninsula, Java, and Sumatra. Mydaus meliceps2. Java and Sumatra. Mustela flavigula. India to China. ,, nudipes. Malay Peninsula and Sumatra. Lutra sumatrana. Malay Peninsula, Sumatra, and Java. ,, cinerea. India to Java and China. RODENTIA. Sciuropterus pulverulentus. Malay Peninsula. ,, horsfieldi. Malay Peninsula and Java. „ setosus. Sumatra. ,, genibarbis. Java. ,, nigripts (Palawan). Pteromys nitidus. Malay Peninsula and islands. „ phceomelas. Rhithrosriums macrotis. Sciurus bicolor. E. Himalaya to Siam and (?) Celebes. „ prevosti. Malay Peninsula, Sumatra, and Celebes. ,, hippurus. Malay Peninsula and islands. pryeri. 1 Jentink, Notes Leyden Museum, vol. XVI. p. 210 (1894). 2 The form from Calamianes has been separated as M. marc Jut. 298 THE ORIENTAL REGION. [CHAP. RODENTIA (cOHt.). Sciurus brookei. „ tenuis. Malay Peninsula and islands, to Siam. ,, lowi. „ jentinki. ,, notatus. Malay Peninsula, Sumatra, Java, etc. ,, insignis. Malay Peninsula, Sumatra, and Java. ,, ho set. ,, everetti. ,, steerei (Palawan). ,, laticaudatus. Malay Peninsula. ,, soricinus, Java and Sumatra. Nannosciurus exilis. Malay Peninsula and Su- matra. ,, whiteheadi, ,, melanotis. Mus infralnteus. Sumatra. ,, muelleri. Sumatra. ,, sabanus. ,, neglectus. Philippines. ,, jerdoni. E. Himalaya, Tenasserim, Java. ,, alticola. ,, ephippium. Sumatra, Philippines. ,, margaretta. ,, raja. ,, ochraceiventer. ,, whiteheadi, ,, bxodon. ,, baluensis. Chiropodomys major. ,, pusillus. Hystrix crassispinis . ,, pumila (Palawan sub-group). ,, muelleri. Sumatra. Trichys gnentheri. VIII.] SUB-REGIONS. 299 UNGULATA. Bos sondaicus. Burma, Malay Peninsula, Java, and Bali. „ moellendorffi ' (Palawan sub-group). Cervus unicolor2, var. Probably introduced. Cervulus muntjac. Tragulus napu. S. Tenasserim to Java and Sumatra. ,, nigricans (Palawan sub-group). ,, javanicus. Malay Peninsula to Cochin China. Sus vittatus. Java, Sumatra, Amboyna, Batjian. ,, verrucosus. Java, Ceram. ,, barbatus*. „ longirostris. Java. Rhinoceros sumatrensis. Assam to Siam, Malay Peninsula, and Sumatra. Tapirus indicus. S. Tenasserim, Malay Peninsula, and Sumatra. Elephas indicus. Probably introduced. EFFODIENTIA. Manis javanica (Palawan sub-group). In this list the genera Nasalis, Trichys, and Rhithrosciurus are peculiar to the group, while Ptilocercus is almost so ; and, includ- ing the latter, there are no less than fifty-one species peculiar to Malaysia. Of these a very large number are common to Sumatra or the Malay Peninsula, or both together, while a smaller number occur in Java. There are but six species common to peninsular India, among which the Indian elephant and the sambar may have been introduced ; but there are ten which are found in the Eastern Himalaya or Assam. The most remarkable among these is the Himalayan water-shrew (Chimarrogale himalayica), which is only 1 Founded on a skull from the island of Busuanga, in the Calamianes, which probably indicates only a race (? introduced) of the buffalo. 2 The so-called C. equinus is regarded by Dr Blanford as not specifically distinct from C. tinicolor. This being so, it is probable that the Bornean form described by Mr C. Hose (Ann. Mag. Nat. Hist. ser. 6, vol. xn. p. -206) as C. brookei is also a variety. 3 Sits ahanobarbus, Huet, from Palawan, and S. calamianensis, Heude, from the Calamianes, are identified by Dr Nehring (Sb. Ges. Nattirf. Berlin, 1894, pp. 190, 191) with this species. 300 THE ORIENTAL REGION. [CHAP. found in the Eastern Himalaya, the mountains north of Burma, and Mount Kina Balu, in North Borneo ; a musk-shrew ( Croridura fuliginosa\ common to the Eastern Himalaya and Borneo, being likewise unknown elsewhere. These two instances alone are sufficient to prove that Borneo must have been in immediate connection with the lands to the north-west within the period during which the living species of mammals have come into exist- ence; while the restriction of the water-shrew to the mountains seems likewise to imply a former lowering of the temperature of the whole region sufficient to enable the creature to pass from the one area to the other, or perhaps rather to have allowed of its existence in the intermediate lowlands, whence it migrated to its present isolated haunts. That Borneo was connected with the mainland during the Pliocene epoch is proved by the occurrence in that island of the Siwalik Mastodon latidens, the tooth figured on page 173 being of Bornean origin. The large number of species common to Borneo, Sumatra, and the Malay Peninsula also shows that these three areas must have been very recently in connection ; but the excessive number of peculiar Bornean forms seems to indicate that the former island, with the adjacent islets, was the first to be isolated. Even so. however, the extraordinarily large percentage of distinctive types is most remarkable. Regarding the relationship of the Palawan sub-group to Borneo, Mr Everett writes that "although the general facies of the mammalian fauna of the sub-group is clearly Bornean, it is to be noted that no species appears to be peculiar to the group as a whole, a fact which suggests the inference that closely connected as Borneo has undoubtedly been with Balabac and Palawan, and isolated as they have been together from the main- land of Asia, there has also been much isolation of Borneo and Palawan inter se" From Sumatra and Borneo, which have so much in common, and in a somewhat less degree from the Malay Peninsula, Java differs very remarkably as regards its mammalian fauna ; a large number of typically Malayan forms being absent, while others as characteristically Indian are present. In the first place, the orangs (Simia), common to Borneo and Sumatra, are absent : and the elephant and tapir are likewise wanting, the former VIII.] SUB-REGIONS. 30 1 certainly existing in a wild state in Sumatra, although it has been considered that its occurrence in Borneo is due to human intro- duction1. Although the Javan rhinoceros (J?. sondaicus}, as we have seen, is common to eastern Bengal, Burma, and Java, its reputed occurrence in either Borneo or Sumatra does not appear to rest upon any solid basis of fact2; while the Sumatran species (R. sumatrensis], which is common to Borneo, is wanting from Java. It has, indeed, been stated3 that certain teeth from the Plistocene of Borneo are referable to the last-named species, but the molars of both kinds are so alike that it is almost, if not quite, impossible to distinguish between them. A noteworthy circumstance is that whereas there is no Siwalik species allied to R. sumatrensis, yet R. sondaicus is almost indistinguishable from the Siwalik R. sivalensis, and is thus proved to be a very ancient Indian type. As another instance of the distinctness of the mammalian fauna of Java from that of Borneo and Sumatra, we may take the case of the banteng (Bos banteng}, which is wanting4 from both those islands, but is present in Java, the Malay Peninsula, and Burma. Again, the genus Hemigale, of which the type species is common to the Malay Peninsula, Sumatra, and Borneo, is quite unknown in Java. That the latter island was directly connected with the mainland is, of course, proved by such species of existing mammals as are common to the two areas ; but if further evidence be needed, it is to hand in the fossil mammals which have been obtained from Pati-Ajam, in Java5. These comprise Elephas trigonocephalus, E. bombifrons, E. clifti, E. namadicus, E. hysudricus, Sus hysudricus, Bos siva- lensis, and Cervus lydekkeri. With the exception of the first and last (which may prove not to be distinct likewise), all these are Indian forms, E. namadicus belonging to the Plistocene, while the others pertain to the Siwalik fauna; and it may be added that 1 Dr Wallace states that Ursus malayanus is absent from Java, but accord- ing to Dr Blanford this is incorrect. 2 See Jentink, Notes Leyden Museum, vol. xvi. p. 231 (1894). 3 Busk, Proc. Zool. Soc. 1869, p. 409. 4 In the Fauna of British India— Mammalia, p. 490, Dr Blanford gives Borneo, and perhaps Sumatra, as part of the habitat of the banteng, but the animal is omitted from fauna of the former by Mr Everett. 5 K. Martin, Sammlungen Geol. Reichsmusetuns in Leiden, vol. IV (1887). 302 THE ORIENTAL REGION. [CHAP. the first three belong to the group of stegodont, or intermediate elephants. In endeavouring to explain the relationship of the Javan fauna to that of the rest of the Malayan sub-region, Dr Wallace1 was first of opinion that Java, which was evidently isolated before Sumatra and Borneo, had a brief land-connection with the Siamese peninsula, independently of those two islands. This view, how- ever, was subsequently abandoned2, and the following hypothesis proposed. From the evidence of certain Tertiary rocks in Java believed to be of Miocene age, it is considered probable that at the epoch in question that island "would have been at least three thousand feet lower than it is now, and such a depression would probably extend to considerable parts of Sumatra and Borneo, so as to reduce them all to a few small islands. At some later period a gradual elevation occurred, which ultimately united the whole of the islands with the continent. This may have continued till the glacial period of the northern hemisphere, during the severest part of which a few Himalayan species of birds and mammals may have been driven southward, and ranged over suit- able portions of the whole area. Java was then separated by subsidence, and these species became imprisoned there ; while those in the remaining part of the Malayan area again migrated northward when the cold had passed away from their former home, the equatorial forests of Borneo, Sumatra, and the Malay Peninsula being more especially adapted to the typical Malayan fauna — which is there developed in rich profusion. A little later the subsidence may have extended further north, isolating Borneo and Sumatra, but probably leaving the Malay Peninsula as a ridge between them as far as the islands of Banka and Billiton. Other slight changes of climate followed, when a further subsidence separated these last-named islands from the Malay Peninsula, and left them with two or three species which have since become slightly modified. We may thus explain how it is that a species is sometimes common to Sumatra and Borneo, while the inter- vening island (Banka) possesses a distinct form3." 1 Geographical Distribution of Animals, vol. I. p. 359. 2 Island Life, p. 360. 3 As exemplified in the case of the birds of the genus Pitta. VIII.] SUB-REGIONS. 303 Although not taking into account the relationship between the fauna of Borneo and that of the Palawan sub-group, this may be accepted as, on the whole, a very probable explanation of the facts of the case. It may be added that while the Javan rhinoceros (£. sondaicus] is, as already stated, closely allied to the Siwalik R. sivalensis, the nearest ally of R. sumatrensis appears to be the extinct R. schleiermacheri of the European Miocene, thus affording one more instance of affinity between the typical Ma- layan fauna and that of the middle Tertiaries of Europe. • Recent investigations on the mammals from some of the small chain of islands lying to the south-west of Sumatra, such as Nias and Sipora in the Mentawi group, to- gether with Christmas Island1, lying still further to Christmas Islands. the south, have shown that they differ very markedly from those of the larger islands. From Sipora, in addition to bats, Mr O. Thomas2 records the following species, of which those peculiar to the island are printed in italics : — Semnopithecus potenziani. Macacus nemestrinus. Widely spread. Tupaia ferruginea, var. hypochrysa. Java. Paradoxurus, sp. Pteromys nitidus. Widely distributed. Sciuropterus lugens. ,, aurantiacus. Banka. Sciurus melanogaster. ,, fraterculus. Mus siporanus. „ raja. N. Borneo. From Christmas Island the same writer3 records a peculiar species of fruit-bat (Pteropus natalis], a variety of a widely-spread musk-shrew (Croddura fuliginosa), and two peculiar rats (Mus macleari, and M. nativitatis], remarkable for their large size. Regarding the bearings of the faunas of these islands on the general problem of distribution, Mr Thomas writes that the collec- 1 This must not be confounded with the island of the same name in Polynesia. 2 Ann. Mus. Civ. Genova, Ser. 2, vol. xiv. pp. 660 — 672 (1895). 3 Proc. Zool. Soc. 1887, pp. 511—514, and 1888, pp. 532 — 534. 304 THE ORIENTAL REGION. [CHAP. tion from Sipora "does not show the very slightest special relationship to Sumatra, and therefore lends weight to the view that the Mentawi chain is the remnant of a long peninsula or island, similar in shape to, but separate from the Malay Peninsula or Sumatra. Further than this I cannot at present go, mainly because we know so little of the small terrestrial mammals of the other islands of the chain, those of the Nicobars being almost unknown, and those of Simalu, Sibiru, and Pagi entirely, while in Nias-and Engafio the collections consist mainly of bats. Still the few indications there are, such as the relations to each other of Pteropus nicobaricus, modiglianii, and natalis, and of Mus siporamis and macleari, show that the mammals, like the other animals, pre- sent a general similarity throughout the chain the whole way from the Nicobars to Christmas Island." Hitherto the Philippine Islands (exclusive of Palawan, Cala- mianes, etc., which are classed with the Bornean sub-re^on.6 group ') have been regarded as forming a portion of the Malayan sub-region ; but the discovery of a very peculiar mammalian fauna in the mountains of Luzon2 clearly proves their right to form a sub-region apart. This mountain fauna, which it is highly probable may also prove to be existent in Mindanao, is evidently a very ancient one, showing certain indi- cations of affinity with that of Australia; while the plain fauna is of a more modern and generally Oriental type. Curiously enough, the indications of affinity with the fauna of Celebes are by no means strongly marked. Unfortunately, there is at present no knowledge of the palseontological history of the mammalian fauna of the group. The following species of mammals, exclusive of bats, have been recorded from this sub-region3; the names of such genera and species as are peculiar being printed in italic type. PRIMATES. Hylobates leuciscus. Macacus cynomolgus. A distinct race4. Tarsius philippinensis1 '. Nycticebus tardigradus. 1 Vide supra, p. 294. 2 See Thomas, Appendix, No. 31. 3 In addition to the paper cited above, see Bourns and Dear, Appendix, No. n. 4 Often separated as M. philippinensis. 5 Meyer, Abh. Mtts. Dresden, 1894, Art. i, p. i. VIII.] SUB-REGIONS. 305 INSECTIVORA. Galeopithecus volans1. Tupaia everetti. CARNIVORA. Viverra tangalunga. Ranges to Celebes. Paradoxurus philippinensis. Also Bornean. Felis bengalensis. RODENTIA. Sciurus philippinensis. „ mindanensis. ,, samarensis. „ cagsi. Nannosciurus concinnus. Chrotomys whiteheadi. Xeromys silaceus. ' j Phlaomys cumingi. ,, pallidus*. Crateromys schatenbergi. Rhynchomys soricoides. Carpomys melanurus. ,, phczurus. Mountains of Luzon. UNGULATA. Batomys granti. Mus luzonicus. \ ,, chrysocomus. \ „ neglectus. ,, ephippium. > Bos mindorensis. Cervus philippinus. ,, alfredi. Sus celebensis, var.! Mountains of Luzon. Mountains of Luzon. Elsewhere only in Celebes. With the exception of the Tarsius, which is now regarded as a peculiar species, the Primates are all wide-ranging forms, as is also the case with Galeopithecus volans, Viverra tangalunga, and Felis bengalensis. A relationship with Borneo is indicated by the Para- doxurus and two species of Mus ; while the pig is typically a Celebean form. The tamarao (J3os mindorensis] has its nearest ally in the anoa of Celebes, but, as mentioned in an earlier chapter, \ilippinensis. 1 The Philippine race has been separated as G. 2 Doubtfully distinct. 3 Equal S. marchei, Huet. 20 306 THE ORIENTAL REGION. [CHAP. it is suggested that the animal in question is really a hybrid between the latter and the Indian buffalo. The two species of deer are small forms allied to the race of the sambar inhabiting Java and Borneo ; the first in the list being uniformly coloured, while the second is spotted with white at all ages. The so-called Cervus marianus of Luzon appears to be inseparable from C. philippinus. It is remarkable that the six genera peculiar to the group all belong to the Muridce, and that five of these are known only from the mountains of Luzon. Moreover it is quite probable that the species from the latter locality referred to the Australian genus Xeromys ought really to be generically distinguished. It is among the Murtdce that evidences of Australian affinities are alone ex- hibited. As these murines have already been noticed on page 277 it will be unnecessary to allude to them further in this place. Next to the peculiar rodents of the mountains, the most remark- able feature about the fauna of the typical Philippines is the absence of such a number of the most characteristic Malayan genera of mammals. Among the Primates the deficiency of orangs (Simia) is perhaps not very remarkable, but the total lack of langurs (Semnopitkecus) and the presence of only a single species of Macacus and Hylobates are most noteworthy. The macaque is, however, distributed over all the islands of the group, and differs from other forms of its species in its extremely light coloration, so that it is scarcely likely to have been introduced by human agency. Of Malayan genera which are absent, special note may be taken of Linsanga, Arctogale, Arctictis, Cynogale, Herpestes, the wild dogs of the sub-genus Cyon, Ursus, Tragulus1, and Elephas. Almost equally well-marked peculiarities are exhibited by the bird fauna. Apart from the tamarao, which has yet to be proved entitled to rank as a valid species, the lowland mammalian fauna of the Philippines is such as might have well reached the group by means of a narrow connection of limited duration with some portion of the Malay countries ; say, for instance, with Borneo by way of Palawan. That such a connection must have been comparatively recent is indicated by the identity of several of the Philippine 1 Represented by T. nigricans in Palawan. VIII.] SUB-REGIONS. 307 species with Malayan forms and the absence of any peculiar genera save Phlceomys. The absence of such a number of Malayan types indicates, however, either that the connection must have been exceedingly brief, or that a large number of species formerly inhabiting the islands have been destroyed by sub- mergence. On the other hand, the presence in the group of a considerable proportion of widely spread continental genera of birds suggests a more free communication with China than at present exists ; such communication having not improbably taken place by way of Formosa. The mountain fauna of Luzon doubt- less indicates an earlier type of colonisation. Note. — Since the foregoing was written two papers have appeared on the Fauna of the Natuna Islands; viz. O. Thomas, Novitates Zool. Vol. n. pp. 26—28, and Thomas and Hartert, I.e. pp. 409 — 429. 2O — 2 CHAPTER IX. THE HOLARCTIC REGION. Characteristics of the Mammalian Fauna — Mammals of the Eastern Holarctic Region — Plistocene Fauna of the Holarctic Region — Geographical Changes since the Plistocene — Western Division of the Region — Arctic Sub -region — European Sub-region — Central Asian Sub-region — Tibetan Sub-region — Mantchurian Sub-region — Mediterranean Sub-region — Kashmir — Ca- nadian Sub-region — Transition Zone. BY far the most extensive of all the zoological regions of the globe is that which is equivalent to the whole of the Palsearctic and the greater portion of the Nearctic region of Messrs Sclater and Wallace, the one to which Dr Heilprin (after a suggestion of Professor A. Newton) applied the name Holarctic. In defining this region, Dr Heilprin separated from it a Sonoran "transitional region " in the Western Hemisphere, and a similar Mediterranean or Tyrrhenian tract in the Eastern. Of these the former is now accepted as a definite region ; but our knowledge of the distribution of species in the Eastern Hemisphere is either too imperfect, or the interdigitation of the two faunas is too complete to admit of the full definition of a Mediterranean region. Accord- ingly, without prejudice as to what it may be possible to ac- complish in this direction in the future, the Mediterranean area is provisionally included in the Holarctic region. It is, however, important to observe that the reservation by Dr Heilprin of the two transitional tracts already named justifies the use of the term Holarctic even if both such tracts be raised to the rank of separate regions ; and there is accordingly no necessity for the adoption of Dr Blanford's term Aquilonian as equivalent to the restricted Holarctic. Used in the sense here indicated, the Holarctic region will comprise all that portion of Arctogasa lying north of the CHAP. IX.] LIMITS AND FEATURES. 309 Sonoran region in America, and of the Ethiopian and Oriental regions in the Old World. The whole area is extra-tropical ; and, as Dr Heilprin remarks, " no other region can compare with the Holarctic in the manifold variety of its physical characteristics. Every form of terrestrial configuration, or condition of soil and climate that may be represented in any other region, is also repre- sented here, and on an imposing scale. From the ice-bound fields of the far north to the burning desert-wastes of Turkestan on the south, and from the deep forest-grown lowlands to mountain- summits soaring thousands of feet above the level of perpetual snow, we pass through all those various gradations of climate which respectively characterise the Frigid, Temperate, and Torrid zones. Densely covered forest-tracts, supporting, as in the north, a sombre growth of pine and other coniferous trees, or, as in the south, a vegetation of almost tropical luxuriance, alternate with broadly open grass or pasture lands (tundras of Siberia, American prairies and plains), which in some cases support over enormous areas only a very scanty vegetation, and in others display a profuse variety of vegetable productions. It is in this region that, in addition to a most bountiful development of desert tracts, we meet with the most elevated table-land (the Central Asian), and, at the same time, with the greatest expanse of lowland on the surface of the globe, the great plain of Siberia and north-eastern Europe." Although the essential unity of the greater portion of the Nearctic and Palaearctic regions has long been fully recognised by the American zoologists, several attempts to bolster up the alleged distinctness of these artificial divisions have recently been made in England1, one proposal being to recognise an Arctic or Boreal circumpolar province cut off from both areas, although this is practically begging the whole question. If I rightly understand his view, Dr C. H. Merriam2, who is an ardent advocate for the zoological unity of the more northern portions of both hemispheres, would distinguish a Boreal circum- polar region common to both hemispheres ; while in America he recognises south of this a Transition region, followed still more to the south by the Sonoran. In the Old World he would have an 1 Appendix, Nos. 28 and 35. 2 Ibid. No. 19, pp. 24, 63. 310 THE HOLARCTIC REGION. [CHAP. analogue of the Sonoran — practically equivalent to the Mediter- ranean sub-region of European writers — but says nothing about an analogue of the Transition; from which I infer that he would use the term Boreal really as practically equivalent to the Holarctic, if the Sonoran and Mediterranean areas were subtracted. In the New World the Boreal, exclusive of the Transitional, is divided into an Arctic and a Coniferous Forest Boreal zone, the latter being frequently spoken of as the Boreal " zone," in contra- distinction to the Boreal circumpolar "region"; the Arctic zone including the tract beyond the limit of trees. The distinction between the Boreal and Transition areas is certainly not of regional value ; and as the term Boreal is used in several senses, it had better give place to the earlier Holarctic. As has been partially indicated in earlier chapters, and as will be more fully noticed in the sequel, there is undoubtedly a marked distinction between the mammals of North America as a whole and those of Europe and northern Asia, but this has been con- siderably exaggerated by including the Sonoran region in the old Nearctic, and is overshadowed by the number of genera and species common to the two areas and unknown elsewhere. Could a Mediterranean region be satisfactorily denned, the homogeneity of the mammalian Holarctic fauna would be still more apparent ; but this, from the great mingling of northern and southern types which has taken place in the Old World, is, I think, impracticable, As has been already mentioned, it is probable that the western and eastern halves of the Holarctic region have never had more than a comparatively small area of communication by way of Bering Strait, and, therefore, the further south we travel in the two areas the more distinct do the faunas become; while only such forms as are capable of withstanding a certain degree of cold have ever been able to cross at all. It may be added that the evidence for the unity of the Holarctic region is by no means solely dependent upon the mammalian fauna, but is supported by many other groups of animals. To take an instance from the insects, I may quote from Mr W. F. Kirby l, who writes as follows : " Had I been dealing with Lepidoptera only, I would certainly 1 Joitrn. Linn. Soc. — Zool. 1873, p. 432. IX.] CHARACTERISTICS OF THE FAUNA. 311 have united Dr Sclater's Palaearctic region and Nearctic region ; for although the species of North American Rhopalocera are seldom identical with those of northern Asia and Europe, still the genera are the same with scarcely an exception, except a few representatives of South American genera, which have no more right to be considered Nearctic species than the similar chance representatives of African1 forms in North Africa or south-west Europe, or of Indian forms in south-east Europe, have to be con- sidered Palaearctic species." On the other hand, North America differs remarkably from the eastern half of the Holarctic region as regards its land molluscs. Thus the Rev. A. H. Cooke2 writes, that " no district in the world of equal extent is so poor in genera, while those which occur are generally of small size, with scarcely anything remarkable either in colouring or form. The elongated land-shells (Clausilia, Bull- minus] so characteristic of Europe are entirely wanting, but a few Bulimulus, of Neotropical origin, penetrate Texas, and from the same sources come a few species of Glandina (as far north as South Carolina), Holospira (Texas), and Helicina" Probably this poverty is largely due to the unsuitableness of the greater part of North America to molluscan life ; aided by the circumstance that land-molluscs are just the creatures that would have been unable to pass over from Asia by way of Bering Strait. The batrachians, again, which differ most remarkably in their distribution from mammals, are not indicative of the unity of the Holarctic region, the American types being very different from those of the Eastern Hemisphere. According, however, to Mr F. E. Beddard3, " the earth-worms offer the best evidence of any group in favour of the Holarctic region." Although during the Plistocene era— even subsequently to the passing away of the glacial period — elephants, rhino- ceroses, and hippopotami abounded over the greater tics ^he part of Europe, while species of the two former groups ranged as far north as Siberia, and macaques 1 The author obviously means Ethiopian. 2 Cambridge Natural History — Mollusca, p. 339 (1895). 3 Appendix, No. 5, p. 80. 312 THE HOLARCTIC REGION. [CHAP. were found in France and England, yet the Holarctic region is now characterised by the absence of all these animals, save a few species of Macacus on its southern borders ; and if a Mediterranean region could be satisfactorily defined, even these, as well as hyaenas and certain other southern types, would likewise be excluded. The fruit-bats constituting the family Pteropodidce. are likewise practically wanting in this region; and Effodientia are quite unknown. On the other hand, carnivores, such as wolves, foxes, bears, martens, weasels, and the glutton, are abundant ; while the rodents are specially represented by such types as the marmots, beavers, voles, and picas ; and the ungulates comprise the bisons, nearly all the sheep, the true goats (absent in the western half of the region), and all the typical deer. Referring in more detail to the peculiar generic types common to the region as a whole, we have, first of all, among the Insecti- vora the typical or true shrews (Sorex] — which belong to that section of the Soricidcz characterised by having the tips of the teeth stained brownish-red — in the main characteristic of this region, although in America they extend southwards into the Sonoran. In the allied family of the Talpida, the mole-shrews ( Urotrichus], which are near relatives of the European desmans, are represented solely by one Japanese, and a second North American species, although the latter is frequently separated generically as Neuro- trichus. The peculiar genera of Carnivora are but few, although certain groups are either confined to, or very strongly represented in, the region. For instance, among the Felidce the true lynxes — which although generally included in the genus Felts, are by some regarded as entitled to form a genus by themselves — are absolutely confined to this region, where they range as far south as Spain. The bears (Ursus], also, are very strongly represented; brown- coloured species being peculiar to the Holarctic area, as is the very distinct polar species to its Arctic portions. The peculiar sea-otter, the sole representative of the genus Latax, has a distri- bution very similar to that of the mole-shrews ; this animal occur- ring on the coasts of Kamschatka and the Kurile Islands, as well as on those of the Aleutians. The wolverene (Gulo}, which is like- wise the only member of its genus, has a more extended range, IX.] RODENTS. 313 being found throughout the forest regions of northern Europe, Asia, and North America, while in the Plistocene era it ranged as far south as England. Although marine mammals are for the most part omitted in this work, the walruses (Trichechidcz), which now have a circumpolar distinction, can scarcely be omitted, since these animals never wander far from land. Remains of the existing forms have been disinterred from the peat of the English fens, while tusks of fossil species have been discovered in the Pliocene Crag of the east of England, and also in the corre- sponding deposits of Belgium. Passing on to the rodents, we have in the Sciuridce the ground- squirrels or chipmunks ( Tamias] practically confined to the region. Although in America they also extend into the Sonoran, in Europe they are unknown in the Mediterranean area. The pouches in the cheeks for storing food and the alternate dark and light longi- tudinal bands down the back serve to distinguish the chipmunks from other squirrels. Fossil remains of the genus, probably belonging to existing species, occur in the Plistocene of Europe and North America. The family of the beavers (Castoridcz) seems always to have been mainly confined to the Holarctic region, one of the two existing species ( Castor fiber] being European, and dating from the English Plistocene, while the other (C. canadensis] is North American. Whereas, however, the latter ranges south- wards into the Sonoran region, the former is unknown in the Mediterranean sub-region. Fossil species of this genus occur in the upper Tertiaries of Europe and North America, where the extinct Chalicomys is likewise met with ; but no beavers are known from either the Siwaliks or the Pikermi beds. Although there are few generic types of Muridce peculiar to the whole region, such as there are are important. Foremost of these are the great tribe of the voles (Microtus1}, constituting the typical representatives of a sub-family nearly allied to the cricetines, but distinguished by the two longitudinal rows of tubercles on the crowns of the molar teeth being modified into alternating triangular prisms, and likewise by these teeth being generally of the hypsodont type. In the Old World they are found from the Arctic zone to Asia Minor, while 1 Commonly known by the later name of Arvicola. 3 H THE HOLARCTIC REGION. [CHAP. in America they enter the Sonoran region. As fossils, they appear to be first known from the upper Pliocene Crag of England ; and were thus probably evolved within the Holarctic region from the more generalised cricetines at a comparatively late epoch. Nearly allied are the lemmings (My odes), which are, however, a more northern type, unknown in the Mediterranean sub-region, and likewise in the Sonoran. Still more northern, and indeed circum- polar in its range, is the banded lemming, which alone represents a genus (Cuniculus) distinguished from the last by the absence of external ears, the short and thick fur on the feet, the rudimentary first toe of the fore feet, and the elongation of the two middle claws of the same. The second of the two families peculiar to the region is that of the picas, or tailless hares (Lagomyidce), comprising small hare -like animals with short ears, of which all the living forms are included in the single genus Lagomys. While the majority of the picas are confined to the highlands of Central Asia, some are found on the first snowy range of the Himalaya, and both south-east Europe and the Rocky Mountains severally possess a representative. Fossil forms are common in the middle and upper Tertiaries of Europe, as far south as Sardinia. Although the hare-tribe (Leporidcz) have an almost cosmopolitan distribution, the majority of species of Lepus are inhabitants of the Holarctic region, Central Asia being especially rich in representa- tives of the genus. Among the Bovidtz, the bisons, which form a well-marked group of the genus Bos, may be regarded as now characteristic of the region, although the American Bos americanus ranges into the Sonoran. In addition to the European B. bison, which was for- merly spread over the greater part of Europe and during the Plistocene extended into Arctic America, there is also the some- what aberrant yak (B. grunniens) of Tibet, In the Plistocene the range of the group was somewhat more extensive, remains of an extinct species having been found in deposits of that age in Texas ; and there is likewise another from the Pliocene of northern India. The sheep constitute a group mainly characteristic of the Holarctic region ; their headquarters being the Central Asian plateau, where they are more numerous than anywhere else in the world, although one species (Ovis vignei) impinges on the IX.] LIST OF THE FAUNA. 315 north-western frontier of the Oriental region, while the single North American form also enters the Sonoran. In a fossil state it is possible that sheep occur in the Indian Siwaliks, but else- where they are first known from the Forest-bed of the Norfolk coast, belonging to the early part of the Plistocene epoch. More nearly allied to the sheep than to the goats, the musk-ox (Ovibos) of Arctic America is now extinct in the Old World, but as it is abundant in the Plistocene of Europe and Asia, where it ranged as far south as England, it is surely entitled to rank among the forms common to the western and eastern halves of the Holarctic region. Nearly allied extinct species occur in the Plistocene formations of the United States. Among the Cervidce. there are three types common to the entire region. The first, or Elaphine group, includes the typical members of the genus Cervus, as represented by the red deer (C. elaphus) of the Old World and the wapiti (C. canadensis) of North America; this group being characterised, among other features, by the general presence of both a brow- and a bez-tine to the antlers, although the latter is wanting in the North African variety of the red deer, and also in the Tibetan C. thoroldi. The alliance between the wapiti and some of the forms inhabiting Central Asia is so close as to render it doubtful whether they are really anything more than varieties of a single species. The single species of elk (Alces) is common to both halves of the region ; and the same is also the case with the reindeer (jRangtfer), which although now not found to the south of Europe, ranged during the Plistocene era into the south of France. The genera or groups which may be regarded as charac- teristic of the entire Holarctic region may be tabulated as follows, those which are practically peculiar being printed in italics : — Insectivora. SORICID^E. Sorex. In America ranges into Sonoran. TALPID^E. Urotrichus. Japan and N. America. Carnivora. FELID^E. Felis, the true lynxes solely Holarctic. MUSTELID.E. Latax. \ One species common to both Gulo. J hemispheres. TRICHECHW&. Trichechus. THE HOLARCTIC REGION. [CHAP. Rodentia. Tamtas. Ranges into Sonoran. Arctomys. Spermophilus. CASTORID^E. Castor. *) .. 0 ,, ,,-. \ Also Sonoran. MURID^E. Microtus. } Myodes. Cuniculus. LA COM YIDSE. Lagomys. LEPORID^:. Lepus, the greater number of species Hol- arctic. Ungulata. Bos, the bison group chiefly Holarctic, al- though the American species reaches the Sonoran. Ovis, just touches Oriental, and also reaches Sonoran. Ovibos, now extinct in the Old World, where it was common in the Plistocene. The arguments for the unity of the Holarctic region are, how- ever, by no means confined to the case of genera, for there are a number of species which either have a circumpolar range, or which are represented by closely allied forms in the opposite hemisphere. It is true, indeed, that many of these are now more or less exclusively Arctic in their distribution, but some range a long distance to the south ; while during the Plistocene epoch this was the case with the majority. The following list includes the more important species which are either common to the eastern and western halves of the region or have representative forms in the two hemispheres ; those which are strictly Arctic having the letter P appended : — 1. Common lynx (Felis lynx). Canadian lynx (F. cana- densis). P. 2. Wolf (Cants lupus). 3. Fox (Cants vulpes). 4. Arctic fox (Cants lagopus). P. 5. Brown bear ( Ursus arctus). IX.] LIST OF THE FAUNA. 317 6. Polar bear ( Ursus maritimus). P. 7. Sea-otter (Latax lutris). 8. Pine-marten (Mustela martes). American marten (M. ameHcana). 9. Weasel (Mustela vulgaris). 10. Wolverene (Gulo luscus). 11. Walrus (Trichechus rosmarus). 12. Arctic vole (Microtus rutilus). P. 13. Common lemming (Myodes lemmus). American lemming (M. obensis). 14. Banded lemming (Cuniculus torquatus). P. 15. European beaver (Castor fiber}. American beaver (C. canadensis]. 1 6. Mountain hare (Lepus timidus^}. 17. Bison (Bos bison). American bison (B. americanus). 1 8. Kamschatkan sheep (Ovis nivicola). Bighorn (O. cana- densis}. 19. Musk-ox (Ovibos moschatus). 20. Central Asian deer (Cervus eustephanus). Wapiti (C. canadensis}. 2 1 . Elk (Alces machlis}. 22. Reindeer (Rangifer tarandus}. If the Plistocene be taken into consideration there may be added the mammoth (Elephas primigenius) and the horse (Equus caballus), remains of both of which have been discovered in Eschscholtz Bay, where those of the European bison also occur. In this list it may be noticed that the North American lynx is so closely allied to the European form that it has been a question whether it is really more than a local variety. Although the American wolf has been separated as a distinct species, it is now generally identified with the European form ; the same being also the case with the so-called cross-fox of North America, which, together with another form from the Himalaya, and a third from North Africa, may be considered a mere variety of the ordinary 1 This name is commonly applied to the English hare, although it properly belongs to the more northern species. The so-called Polar hare (Z. glacialis) of Arctic America appears to be only a variety. 3l8 THE HOLARCTIC REGION. [CHAP. fox. Much discussion has taken place with regard to whether any of the bears of North America are distinct from the common brown bear ( Ursus arctus), with its many Asiatic varieties. According, however, to one of the latest memoirs on this subject1, all these forms appear mere varieties of the latter, the so-called cinnamon and black bears presenting more distinctly marked differences than does the grizzly. The American marten is so nearly related to the European marten and the Asiatic sable that it is almost impossible to point out valid characters by which the three forms •can be specifically distinguished. In the case of the walrus, the Pacific form is distinguished by certain external features from the one inhabiting the Atlantic coasts, although these are scarcely of sufficient importance to be ranked as specific. In regard to the Arctic vole, it may be mentioned that although it is typically a polar form, yet it is represented in southern Europe by the bank- vole (Microtus glareolus) and in the United States by Gapper's vole (M. gapperi\ both of which may be regarded as southern climatic offshoots from the northern stock. Among the other species of rodents it will suffice to mention that the European and American beavers are merely distinguished from each other by the relative lengths of the nasal bones of the skull. And it may be added that the Kamschatkan wild sheep is so closely related to one race of the bighorn, or Rocky Mountain sheep, that it is very questionable whether the two are really entitled to specific distinction. The same is also the case with the two deer mentioned in the list. Reference has already been made to the circumstance that the musk-ox is extinct in the eastern division of the region. A few of the American forms, such as the bear, beaver, bison, and bighorn, enter the limits of the Sonoran region. Although, as will be shown immediately, there are a large number of generic types respectively confined to its eastern and western divisions, the lists given above, especially the one relating to the species, are amply sufficient to demonstrate the essential unity of the Holarctic region. None of the other zoological regions have anything like the number of common or representa- 1 A. E. Brown, Proc. Ac. Philad. 1894, pp. 119, 129. Merriam, P. BioL Soc. Washington, vol. x. pp. 65 — 83 (1896), regards the N. American bears as forming several distinct species. IX.] EASTERN DIVISION. 319 tive species which characterise the two divisions of the present one. It must, moreover, be remembered that in the case of the other regions we have taken the whole of the peculiar generic types into consideration, although many of them are confined to small portions of such regions, whereas in the present instance only those which range over nearly the whole area have been mentioned. If, for instance, we were to take the genera respec- tively confined to the Indian and Malayan areas of the Oriental region, it would be found that the distinction between the two areas would be nearly or quite as marked as are the two great divisions of the Holarctic, while in the matter of species the differences between the two would be far greater. There would accordingly be stronger grounds for making an Indian and a Malayan region than there are for the separation of a Palsearctic and a Nearctic. Having now discussed the leading mammalian types character- istic of the Holarctic region as a whole, it remains to notice those confined to its eastern division, t °f after which such as are restricted to its western half Holarctic will be taken into consideration. And here it is advisable to repeat that since the communication between eastern Asia and North America by way of Bering Strait appears always to have been of very limited extent, as we proceed south in the two great continental areas the difference between their faunas appears more and more strongly marked. Accordingly, if it were possible definitely to establish a Mediterranean region, the dis- tinction between the faunas of the eastern and western divisions of the Holarctic would be much less than is the case under the arrangement here followed. Passing by the bats entirely, no notice will be taken of groups which, like the hedgehogs (Erinaceus), are spread over several of the regions of Eastern Arctogaea, since these are not in any way characteristic of the eastern division of the Holarctic region1. The first mammal that presents itself for notice is accordingly the water-shrew, which is the sole representative of the genus Crosso- pus, and belongs to that division of the Soricidce in which the 1 Such types have been discussed when considering the fauna of Eastern Arctogaea, supra, p. 181. 320 THE HOLARCTIC REGION. [CHAP. teeth are stained red ; the especial characteristic of the genus being the thickly-fringed feet and tail. The water-shrew is a typical Holarctic mammal, ranging as far east as the Altai mountains and unknown in the Mediterranean sub-region. In the second division of the same family, or that in which the teeth are white, there is a peculiar shrew from the Kirghiz steppes nearly allied to the FIG. 66. RUSSIAN DESMAN (Myogale moschata}. widely-spread musk-shrews (Croridura), but constituting a genus (Diplomesodon) by itself. To the same sub-family belongs the Tibetan water-shrew (Nectogale], which is likewise the solitary representative of its genus, and is closely allied to Chimarrogale, which has, as elsewhere stated, one Oriental and one Japanese species. Both these types are accordingly closely connected with the Oriental region, although nothing is known of their past history. Far more characteristic of the eastern half of the region IX.] EASTERN DIVISION. 321 under consideration are the two species of desman (Myogale), which are aquatic insectivores, with long trunk-like snouts, some- what intermediate between the shrews and the moles. Of the two living species, the smaller is confined to the region of the Pyrenees, extending as far north as the Department of the Landes, while the other is now restricted to south-eastern Russia, although its fossilised remains have been discovered in the Plistocene Forest- bed of the east of England. Extinct species occur in the Mio- cene and upper Oligocene of the Continent. A slate-coloured insectivore, with the external form of a shrew but the skull of a mole, inhabiting Eastern Tibet, constitutes the genus Uropsilus ; and the more mole-like creature known as Scaptonyx is likewise from the same locality. In any case, these two animals only just enter the border of the region, so that they cannot be regarded as characteristic Holarctic types ; and, indeed, the Moupin district of Eastern Tibet is included by Dr Wallace in the Oriental region, although it is assigned by others to the Holarctic. Although two species occur to the south of the Himalaya, the true moles (Talpa) are very characteristic of the eastern Holarctic region, the common species ranging from England to Japan, and dating from the epoch of the Norfolk Forest-bed; while fossil species, some of which have been separated as Protalpa, range through the Tertiaries of Europe to the epoch of the upper Oligocene. By some writers Talpa moschata of Eastern Tibet is distinguished as Scaptochirns. Among the Carnivora the widely spread Old World genera Genetta, Herpestes, and Hyczna enter the Mediterranean sub-region, but are unknown elsewhere in the Holarctic area at the present day. The Ursidce include a most remarkable generic type known as ALluropus, represented by the parti-coloured bear of Tibet, in which the cheek-teeth present a decided approximation to the extinct Hytznarctus and also resemble those of the panda (jElurus). This genus is another of the border-forms between the Holarctic and Oriental regions. On the other hand, the badgers (Meles) are very characteristic of the area under consideration, ranging from England through the rest of Europe to Japan and China, where one species enters the Oriental region, being found as far south as Hongkong. Remains of extinct badgers occur in the lower Pliocene of Persia. L. 21 322 THE HOLARCTIC REGION. [CHAP. The list of more or less peculiar generic types among the rodents is relatively large. Foremost among these stands a very large flying-squirrel {Eupetaurus}, inhabiting the regions north of Kashmir and Tibet, and distinguished from all other members of the family to which it belongs by its tall-crowned (Jiypsodonf] cheek-teeth. In the dormouse family (Myoxida) the squirrel- tailed species, which is the sole representative of the genus ~Myoxus in its restricted sense, and the common dormouse, alone constituting Muscardinus, are exclusively European ; fossil forms occurring through the upper and middle Tertiaries. In the Muridce. the hamsters (Crtcetus), if regarded as generi- cally distinct from their allies the white-footed mice (Sitomys] of the New World, are absolutely characteristic of the eastern division of the Holarctic region, where they range over a large portion of Europe and northern Asia. Extinct species are abundant in the European Tertiaries. Two genera of the mole-like rodents, having the dentition of voles, but approximating in the form of their body and limbs to the moles, constitute a sub-family which is also restricted to this area. Of these mole-voles, the first genus (Ellobius] is represented by one species from Russia and another from Afghanistan, while the second (Siphneus) includes several forms from North and Central Asia; fossil species of the latter having been described from the Plistocene of the Altai and the Pliocene of northern China. The great mole-rat (Spalax typhlus) of southern Europe, Persia, Mesopotamia, Syria and Egypt, repre- sents the only genus among the Spalacidce. which is confined to the present area. The Dipodidce, on the other hand, contain several characteristic eastern Holarctic generic types. The most aberrant of these are the rat-like animals constituting the genus Sminthus, of which one species inhabits eastern and northern Europe and Central Asia, while a second is found in Kashmir, and a third in the Kansu district of China. Of the more typical forms, the true jerboas (Dipus), characterised by having only three toes, are exclusively Holarctic, ranging from Egypt into Central Asia, where they always frequent desert districts. Of the genera with five toes, Euchoretes is represented by a single long-snouted and long-eared species from the neighbourhood of Yarkund ; and Platycer corny s, which differs by its flattened and lancet-shaped IX.] EASTERN DIVISION. 323 tail, includes several species, extending from Siberia to Nubia, so that the genus just enters the Ethiopian region. It is repre- sented in the Plistocene of northern Asia. Lastly, there is the genus Alactaga, of which there are several species, mostly inhabit- ants of Northern and Central Asia. The typical A. jaculus extends, however, into Persia and southern Russia, while in the Plistocene it ranged as far west as Germany ; and another species is an inhabitant of Afghanistan. Although mainly an Ethiopian FIG. 67. HEAD OF SPANISH IBEX (Capra pyrenaica). and Neotropical family, the Octodontidce. have an Holarctic repre- sentative in the gundi (Ctenodactylus\ which inhabits the borders of the Sahara in the neighbourhood of Tripoli. The extinct Pellegrinia, of the Sardinian Pliocene, also belongs to the same family. 21 — 2 324 THE HOLARCTIC REGION. [CHAP. Among the artiodactyle ungulates there are six genera, either entirely or mainly confined to the eastern Holarctic region. Of these, the true goats — that is to say, the members of the genus Capra, as distinct from the Oriental and Arabian Hemitragus — are almost exclusively Holarctic, although C. walie inhabits the Abyssinian highlands, and C. siniatica, of Palestine and upper Egypt, may also enter the confines of the Ethiopian region. All the goats, it may be observed, are essentially mountain-dwelling animals, and the occurrence of the same species of ibex (C. sibiricd) in both the Altai and Himalaya is a clear proof of the former prevalence of colder conditions, as without these the ani- mal could not have passed from the one range to the other. The sheep also — in spite of the existence of outlying North American species, and a variety of one Central Asian form (Ovis vignei] which enters the north-western confines of India — are mainly cha- racteristic of the area under consideration, attaining their great- est numerical development, and also their maximum size, in the highlands of Central Asia. The range of the genus includes almost the whole of the eastern Holarctic region, the mouflon (O. musimon) inhabiting the Corsican islands, and the aberrant arui (O. tragela- phus] being found in northern Africa. It is possible that fossil sheep occur in the Indian Siwaliks (where remains of a goat allied to the Himalayan markhor are also met with), and a large species is definitely known from the Norfolk Forest-bed. The next on our list is the remarkable goat-like antelope from the hills to the north of the Assam known as the takin (Budorcas], which is allied to the Oriental Nemorhczdus, and is therefore probably an immi- grant into the region from the southward. Allied to this group is the chamois, or gemse (Rupicapra), now confined to the higher mountain-ranges of Europe from the Pyrenees to the Caucasus, but which during the Plistocene epoch ranged over many of the lowlands. Among the true antelopes, the addax (Addax], an ally of the oryx group, is an exclusively Mediterranean type, inhabiting North Africa and Syria, where there are also representatives of other genera which are typically Ethiopian. More thoroughly Holarctic are the two peculiar but allied genera Saiga and Panthol- ops, each represented by a single living species. The saiga is now confined to the steppes of western Asia and Eastern Europe, but IX.] EASTERN DIVISION. 325 during the Plistocene epoch extended as far westwards as Germany, France, and England. The chiru, as the representative of the second genus is called, is, on the other hand, an exclusively Tibetan form ; and it is believed that a fossil species occurs in the later Tertiary formations of the same area, where, curiously enough, a rhinoceros also existed. Although gazelles (Gazella) have repre- sentatives in both the Oriental and Ethiopian regions, they are mainly characteristic of the desert districts of the eastern Holarctic region, being especially numerous in North Africa, Syria, and parts FIG. 68. MUSK-DEER (Moschus moschiferus). of Central Asia. An inhabitant of the cooler regions of Asia, where it extends from the south of Siberia to Kashmir and Cochin China, the musk-deer (Moschus) represents a peculiar sub-family of the Cervidoe, confined to the region under consideration. A second species has been described from Kansu, in north-western China, and it is not improbable that the genus is also represented in the Indian Siwaliks. Although agreeing with the musk-deer in the absence of antlers and the presence of long tusks in the upper 326 THE HOLARCTIC REGION. [CHAP. jaw of the males, the Chinese water-deer (Hydropotes), from the valley of the Yang-tsi-kiang, belongs to the more typical Cervidce. Another genus (Elaphodus) more nearly allied to the muntjacs is also Asiatic, being represented by one species from near Ningpo, in China, and by a second from Moupin, in eastern Tibet. Of the true deer (Cervus) there are two groups confined to the area under consideration. Firstly, there is the elaphurine group, repre- sented solely by the aberrant David's deer (C. davidianus\ of northern China; and, secondly, we have the damine group, of which the fallow deer (C. dama) of the Mediterranean countries and the Persian fallow deer (C. mesopotamicus] are the living forms. Allied types occur in the East Anglian Forest-bed, and the gigantic extinct Irish deer (C. gigantens] must likewise be included in the group, all the members of which have the antlers more or less palmated. As regards the camels (Camelus), there is some difficulty in arriving at a satisfactory conclusion, since although a feral race of the Asiatic C. bactrianus is met with in the deserts bordering Kash- gar, it is now pretty well ascertained that really wild camels exist nowhere in the world. Still, however, as fossil species of the genus are met with in the Pliocene of the Siwalik Hills (on the borders of the Oriental and Holarctic regions) and in the Plisto- cene of Algeria, it is probable that the group is of Holarctic origin1. The foregoing survey may be summarised as follows : the names of such genera or groups as are mainly or exclusively con- fined to the eastern division of the Holarctic region being printed in italic type. Insect! vora. SORICID^:. Crossopus. Diplomesodon. C. Asia. Nectogale. Tibet. TALPID/E. Uropsilus. \ „ _.. _, -V \ E. Tibet. Scaptonyx. ) Talpa. Also enters Oriental. URSID^E. sEluropus. Tibet. MUSTELID^E. Meles. Enters E. Oriental. 1 See page 281. IX.] EASTERN DIVISION. 327 Rodentia. MURID/E. SPALA CID&. OCTODONTID^E. Ungulata. CERVID^E. Eupetaurus. Tibetan. Myoxus. Muscardinus. Cricetus. Ellobius. Sip /incus. Spalax. Sminthus. Dipus. Euchoretes. Central Asian. Platycer corny s. Enters Ethiopia. Alactaga, Ctenodactylus. North African. CAMELID/E. Capra. An outlying Ethiopian species. Ovis. One N. American species, and one from Central Asia entering Oriental. Budorcas. Tibetan. Rupicapra. European. Addax. Mediterranean. Saiga. Central Asian. Pantholops. Tibetan. Gazella. A large proportion of the species E. Holarctic. Cervus. The Elaphurine and Damine groups exclusively E. Holarctic; the former Central Asian, and the latter Mediterranean. Elaphodus. E. Tibet and China. Hydropotes. Chinese. Moschus. Asiatic. (?) Camelus. With the possible exception of the Camelidce, none of the families in the foregoing list are peculiar to the area in question— a feature presenting a marked contrast to the lists of the character- 328 THE HOLARCTIC REGION. [CHAP. istic mammalian genera of the Ethiopian and Oriental regions given above. The total number of genera which can in any sense be considered as peculiar to the eastern half of the Holarctic region does not exceed thirty ; and among these Uropsilus, Scapt- onyx, Elaphodus, and Hydropotes can only be regarded as intruders from the Oriental region ; while Ctenodactylus and Addax are manifestly Ethiopian types. Indeed, if a Mediterranean region were established, the whole of these, and probably also the true Tibetan forms, would have to be removed from the lists. Apart from the absence of peculiar families, the list bears no comparison as regards numbers with that of the mammalian genera distinctive of the Ethiopian region ; while, with the aforesaid deductions, it is also considerably inferior to that of the Oriental region. All this confirms the conclusions already drawn as to the inad- visability of regarding the area under consideration as a separate zoological region. The Pliocene and earlier Tertiary faunas of this area having already been considered in connection with Eastern Arctogsea in general in an earlier chapter, we may Eastern Hoi- pass On to a brief review of the Plistocene mammals arctic Region. of the eastern division of the Holarctic region, pre- paratory to the consideration of the sub-regions into which the latter is divided. The Plistocene period may be taken in England to commence with the Forest-bed of the Norfolk coast, which overlies the topmost of the Pliocene Crag series, and is itself overlain by the glacial deposits. To a later epoch of the same period belong the brick-earths and gravels of our river valleys, as well as the cavern-deposits ; many of these being either of post- or inter-glacial age. During the Plistocene period two very remarkable differences from the existing state of things have to be noticed. In the first place, the eastern Holarctic region was at that time very much less distinctly differentiated from either the Ethiopian or the Oriental than is the case at the present day ; macaques, hyaenas, the lion, rhinoceroses, hippopotami, and elephants abounding in Europe even as far north as England. The second point is the curious mixture of remains of existing species of mammals respec- tively characteristic of hot and cold climates met with in many IX.] EASTERN PLISTOCENE FAUNA. 329 parts of England and France. For instance, the glutton, rein- deer, Arctic fox, and musk-ox are animals whose presence seems indicative of a more or less decidedly Arctic climate ; many of the voles, picas, jerboas, and susliks, together with the saiga antelope, appear to point with equal force to the prevalence of steppe-like conditions ; while the hippopotamus and spotted hyaena seem as strongly in favour of a subtropical state of things. Nevertheless, remains of several of these groups have been found in such close association as to leave no doubt that the animals lived and died hard by where they are now buried. Much evidence on this point has been collected by Sir H. H. Howorth1, who writes as follows : " Cuvier, whose prejudices were the other way, was long ago con- strained to write of the remains of reindeer found with those of the mammoth and rhinoceros in the cave at Breugue : ' II ne faut pas douter qu'il [the rhinoceros] n'ait dte enseveli avec lui [the reindeer] a Breugue. Ses os y e'taient pele-mele avec ceux de ce grand quadrupede, enveloppes dans la meme terre rouge, et revetus en partie de la meme stalactite.' In the high-level gravels of the Thames valley the mammoth and woolly rhinoceros occurred with the Elephas antiquus ^ while in the low-level gravels the Rhinoceros leptorhinus and the hippopotamus occurred with the bison and the musk-ox2, together with a worked flint. The lemming and the reindeer occurred with the lion and hyaena at Bleadon, the lemming with the lion and hyaena at Wookey-Hole. The reindeer and the grizzly bear were associated with the hippopotamus at Cefn, and the E. antiquus with the mammoth at Durdham Down. The hippopotamus and the E. antiquus have been found with the reindeer and bison in Kirk dale Cave ; the hippopotamus with the wild boar, the rein- deer, the mammoth, and the E. antiquus at Brentford. Lartet says that in France remains of the hippopotamus have been found in one cave, that of Arcy, in which the reindeer has also occurred, accompanied by a worked flint. At St Acheul and in the Somme valley the same two animals have occurred together, and also at Levallos, in the Seine valley. At Viry Noureuil, near Chauny 1 The Mammoth and the Flood (London, 1887), pp. 114, 115. 2 The author uses the term musk-sheep for this animal ; a few other verbal alterations have been made in the quotation. 33° THE HOLARCriC REGION. [CHAP. (Aisne), the mammoth and Rhinoceros antiquitatis have occurred with the hippopotamus, the reindeer, and the musk-ox. At Bicetre, close to Paris, the lion is associated with northern voles, a marmot, a lizard, and a snake. At Montmorency the mole and the hedgehog have occurred with the hamster, the suslik, and the pica. In Auvergne, M. Pomel has found an elephant and the woolly rhinoceros with a cat, a suslik, and a hamster, together with snakes, lizards, frogs, and with shells such as are still found in the district. In Germany it is the same. At Westeregeln the lion and the spotted hyaena, the mammoth and rhinoceros were found with the marmot, the suslik, the lemming, the pica, and the rein- deer. At Thiede, the mammoth, woolly rhinoceros, the horse, the ox, the reindeer, the Arctic fox, the lemming, and the pica; and so we might continue throughout the majority of the German caverns." Many attempts have been made to reconcile these apparently contradictory facts ; one of the older views being that while the tropical types of mammals lived during warm interludes, they migrated southwards with the incoming of colder conditions to give place to the more Arctic fauna. The associations mentioned above render it, however, perfectly certain that such an explanation is not the right one. On the other hand, it must be remembered that there is yet much to be learnt about the effects of climate on mammals ; and the mammalian fauna of the Tibetan plateau shews that many types of animals formerly regarded as more or less essentially tropical or sub-tropical are capable of withstanding a winter climate of extreme severity. Thus in parts of Tibet, as well as in Kashmir, langurs and macaques may be seen leaping among the snow-clad branches of pines. Still it is perhaps diffi- cult to understand how two such animals as the hippopotamus and the reindeer could have inhabited the same locality contem- poraneously1. In spite of this association of Arctic and sub-tropical forms, there appears, however, to be evidence of a northern and southern 1 The present writer is not prepared to accept the view of Mr A. H. Keane (Mthnology, Cambridge Geographical Series, p. 65, 1896) that the reindeer has only recently become adapted to a northern habitat. Among other circum- stances, its remains are unknown from the Forest-bed. IX.] EASTERN PLISTOCENE FAUNA. 331 type of Plistocene fauna, England being apparently somewhere near the border-line where the two met, and, at times at least, overlapped each other. Probably all or nearly all of the living European mammals were in existence at the same time ; but most of these need not be referred to here, attention being concentrated on those which are either extinct, or are now inhabitants of other regions or districts. The following list includes the more im- portant of these forms ; those which are decidedly northern types being indicated by a *, and those which appear essentially southern with a f. The scientific names of extinct species and genera are printed in italic, and of those still living in ordinary type. Primates. t Barbary Ape. Macacus inuus. f ,, ,, pliocenus. England. f ,, „ suevicus. Switzerland. f ,, ,, tolosanus. France. Carnivora. fLion. Felis leo. f Kafir Cat. Felis caffra. t Sabre-tooth. Machcerodus latidens. f Spotted Hyaena. Hyaena crocuta. t Striped Hyaena. ,, striata. Cave Bear. Ursus spelceus. * Arctic Fox. Canis lagopus. European Wild Dog. ,, (Cyon) europ&us. „ Hunting Dog. Lycaon anglicus. England. * Wolverene. Gulo luscus. Rodentia. t Maltese Squirrel. Leithia melitensis\ * Giant Beaver. Trogontherium cuvieri, * Northern Vole. Microtus rutilus. Sardinian Pica. Lagomys sardus. 1 Originally described as a gigantic dormouse, but shown by the writer in a communication to the Proceedings of the Zoological Society, 1895, p. 860, to be allied to the Sciurida. 332 THE HOLARCTIC REGION. [CHAP. Ungulata. Aurochs. Bos taurus, var. primigenius. *Musk-Ox. Ovibos moschatus. t Barbary Sheep. Ovis tragelaphus. f Spanish Ibex. Capra pyrenaica. English Gazelle. Gazella anglica. t Saiga Antelope. Saiga tartarica. Irish Deer. Cervus giganttus. f Hippopotamus. Hippopotamus amphibius. fPentland's Hippopotamus. ,, pentlandi. * Woolly Rhinoceros. Rhinoceros antiquitatis. Megarhine ,, ,, megarhinus. fLeptorhine ,, ,, leptorhinus. t Etruscan ,, ,, etruscus. * Elasmothere. Elastnotherium sibiricum. * Mammoth. Elephas primigenius. t Straight-tusked Elephant. „ antiquus. t Southern Elephant. „ meridionalis. ( ,, melitensis. \ Maltese T Dwarf Elephants. ._ . . \ , . , ( „ mnaidnensts. I Islands. t African Elephant. ,, africanus. In this list the Barbary ape is now confined to North Africa and Gibraltar. The lion, although now restricted to Africa, India, Persia, and Mesopotamia, ranged during the historic period into Thessaly ; while the Kafir cat is solely African. The sabre- toothed tiger of the caves was the last survivor of a genus common in the Pliocene, which in the Plistocene is unknown further north than Cromer. The spotted hyaena, whose remains are so abundant in the English caves, is, as we have seen in an earlier chapter, now restricted to southern Africa ; while the striped species, which dates from the upper Pliocene, now ranges from north Africa to India. The cave-bear, a gigantic extinct species, was distinguished from the brown bear by the more complex structure of its molar teeth. Of the Canidce, the European wild dog has its nearest living ally in the Altai, the other species of the group being Oriental ; while the European hunting-dog, which is known only by a single jaw from the Glamorganshire caves, appears to have been closely IX.] EASTERN PLISTOCENE FAUNA. 333 related to the living Cape species. Among the rodents, it is only necessary to mention that the giant beaver (Trogontheriuni) represents a distinct genus ranging from the Norfolk Forest-bed to Siberia; and also that the Maltese squirrel (Leithia) was restricted to the islands from which it takes its name. In the ungulates, the aurochs l was the gigantic ancestor of the domestic cattle of the present day, but is unknown living in a wild state. The arui, or Barbary sheep, is now restricted to north Africa; while the Spanish ibex is confined to the mountains of the Iberian penin- sula, its fossil remains occurring in the Gibraltar caves. The Irish deer, distinguished by its great size and widely-spreading antlers, was an ally of the fallow-deer, with which it is connected by means of another extinct species or variety (C. ntffi); and it may be mentioned that there are species of extinct deer from the Forest-bed, which it is unnecessary to name in this place. The latter deposits are the source of the known remains of the English gazelle. The common hippopotamus, which dates from the upper Pliocene of Italy, is now exclusively confined to Ethiopian Africa, but in the Plistocene is known to have wandered as far north as Yorkshire. Pentland's hippopotamus is a smaller species from Italy and the Mediterranean islands, where there may be a second still smaller form. Of the rhinoceroses, R. antiqiiitatis, which is exclusively Plistocene, ranged from Central Europe to Siberia ; its remains being dredged abundantly, in common with those of the mammoth, from the Dogger Bank, in the North Sea. The relation- ship of this species to the extinct Indian Rhinoceros platyrhinus and the living African R. simus has been alluded to in a previous chapter. The other three European species of the genus, which, like the last, were two-horned and devoid of front teeth, date from the Pliocene, and form a group differing remarkably in dental characters from R. antiquitatis. While two of these species were southern types, the third accompanied the mammoth and woolly rhinoceros in Siberia. A near ally of the rhinoceroses, the huge Siberian Elasmotherium, differed remarkably in the structure of its cheek-teeth, which are tall-crowned, and shew some approximation to those of the horses. 1 The European bison is frequently miscalled the aurochs. 334 THE HOLARCTIC REGION. [CHAP. From a distributional point of view, the European Plistocene elephants are of especial interest. Foremost and best-known of these is the mammoth (Elephas primigeniiis], which, as stated in an earlier chapter, was a very near ally of the existing Indian species, although distinguished — as we know from the evidence of specimens preserved in the frozen soil of Siberia — by its coat of woolly red hair, among which were intermingled long bristly black hairs. Curiously enough, traces of this woolly coat have been detected in the Indian elephant, so that it is probable that this species originated in some part of Asia where the climate is colder than is that of India. Regarding the range of this species, Professor Boyd Dawkins1 remarks that "the mammoth is very abundant in the caverns and river-deposits of Britain and of France, and is known to have ranged over the Pyrenees into Spain, from the discovery of specimens in the zinc-mines of Santander. It has been proved by Prof. E. Lartet and Dr Falconer to have lived in the neighbourhood of Rome when the volcanoes of central Italy were active, and poured currents of lava and clouds of ashes over the [site of the] imperial city. It is common in northern and southern Germany, but it has not been found in Europe north of a line passing through Hamburg, or in any part of Scandinavia or Finland. It occurs in the auriferous gravels of the Urals ; and in Siberia, as is well known, it formerly existed in countless herds, being buried in the morasses in large numbers, in the same manner as the Irish elks at the bottom of the Irish peat-bogs. The admirable preservation of some of the carcases is undoubtedly due to their having been entombed directly after death, and then quickly frozen up, a process which need not necessarily imply climatal conditions unlike those of the present time in Siberia." That the mammoth ranged across Bering Strait into Arctic America, is proved by the discovery of its remains in the frozen soil of Eschscholtz Bay; but in the greater part of North America it was replaced by the closely-allied E. columbianus. In eastern Europe there existed a variety or species known as E. armeniaais, of which the molar teeth still more closely resemble those of the Indian elephant than do those of the typical form. The straight- 1 Early Man in Britain (London, 1880), p. 106. IX.] EASTERN PLISTOCENE FAUNA. 335 tusked elephant (E. antiquus) is a more southern Plistocene type, of which the molars are to a certain extent intermediate between those of the living Indian and African species. Still more southerly in its distribution is the gigantic southern elephant (E. meridionalis), of which the remains are found in the upper Pliocene of Italy, as well as in the Plistocene Forest-bed of Norfolk, and equivalent strata at Dewlish, in Dorsetshire. The Maltese Islands were the habitat during the Plistocene epoch of the two or three species of dwarf elephants, which appear to have been nearly allied to the existing African species, but whose size was diminished by the smallness of the areas where they flourished. Lastly, the African elephant, which is now restricted to Ethiopian Africa, has left evidence of its existence during the Plistocene epoch in Algeria, Spain, and Sardinia. The fauna of the Forest-bed period, among which the mam- moth, megarhine rhinoceros, and Irish deer are wanting, is, as already stated, of pre-glacial age, and, on the whole, indicative of a fairly warm climate, although there is some evidence that the musk-ox then ranged as far south as England. At the close of this epoch, the southern elephant, together with a small bear known as Ursus arvernensis, appear to have become extinct. Soon after, glacial conditions made their appearance, causing much dis- turbance and migratory movements among the original southern pre-glacial fauna, and bringing an incursion of northern forms like the reindeer, Arctic fox, wolverene, and musk-ox, as well as of species from the eastern steppes such as the Saiga antelope and the Kirghiz jerboa (Alactagd), together with mountain animals like the chamois, the ibex, and the marmot, into the lowlands of south- western Europe. Among the northern forms that then spread themselves southward were the mammoth and the woolly rhino- ceros, which at this epoch appear to have attained their maximum development. Unfortunately, there is much uncertainty as to the part played by the glacial epoch in the extermination of the large mammals characterising Plistocene Europe. By most English geologists the brick-earths of the Thames valley, which contain remains of rhino- ceroses and elephants in abundance, as well as those of monkeys more sparingly, are regarded as of post-glacial age ; but Prof, von 336 THE HOLARCTIC REGION. [CHAP. Zittel1 considers them pre-glacial, or more probably inter-glacial. If they are either inter- or post-glacial, it is clear that the cold was not the exterminating cause ; and it is quite possible that many or all were killed off by man, although this could scarcely be the case with the Siberian fauna. Be this as it may, there is good evidence that when northern forms, such as the reindeer, wolverene, and banded lemming, had once obtained an entrance into central and southern Europe, they remained there for a considerable time, since they were present during the latter portion of what is known as the palaeolithic epoch. With the advent of the present climatic conditions came in the present woodland fauna of central Europe, constituting what has been termed the squirrel- or bison-epoch ; and from that date, when animals became domesticated, man has exercised a large influence on the fauna. It is important to notice that, in spite of the mingling of northern and southern types in England, France, and Germany, to which allusion has already been made, there seems to have been a distinction between the northern and the Mediterranean fauna throughout the whole of the later Plistocene epoch, such forms as the Barbary sheep and the fallow-deer being essentially southern, although the hippopotamus, as we have seen, extended as far north as Yorkshire. Although the later Plistocene fauna was spread not only over Europe, but also through North and Central Asia, a number of the characteristic European types, such as the hippopotamus, ibex, chamois, fallow-deer, cave-bear, and wild cat, were wanting in Asia. In that area forms are met with which are still character- istic of the same districts. As examples may be noticed : the Mongolian gazelle (Gazella gutturosd], the Himalayan ibex (Capra sibirica), the Persian wild goat (Capra cegagrus), the argali (Ovis argali), the musk-deer (Moschus moschiferus], the tiger (Felts tigris] — of which the remains have been found even within the Arctic Circle — together with a number of smaller forms, such as Siphneus aspalax, Ellobius talpinus, Spalax typhlus, Sminthus vagans, Tamias asiaticus, and Mustela zibellina. Here, then, 1 Appendix, No. 36, p. 189. IX.] LAND CHANGES. 337 are clear indications of a Central Asiatic sub-region as far back as the Plistocene epoch. The foregoing brief survey of the Plistocene mammals of the eastern division of the Holarctic region enables certain deductions to be drawn as to geographical changes which have taken place in the area since the Plistocene- that epoch. In the first place, the occurrence of remains of the tiger in the New Siberian, or Liakov Islands, lying far within the Arctic Circle, indicates the union of those islands with the Siberian main- land ; and this greater extension of the land at the north-eastern extremity of Asia would naturally lead to the conclusion that there was also a land-connection with Alaska across Bering Strait. That such was really the case is proved by the discovery during the voyage of H. M. S. " Blossom," in the years 1825-28, of remains of the horse, mammoth, and bison, in the frozen soil of Eschscholtz Bay, Kotzebue Sound, Alaska1; this evidence being confirmed by the occurrence of the musk-ox in the European Plistocene, as it is also by the number of species of mammals still common to the more northern parts of the two hemispheres. And here it may be mentioned that, according to the researches of the Russian geologists, Siberia, instead of being covered like northern Europe with a continuous ice -sheet during the glacial epoch, had only a number of comparatively small glaciers, so that the pre- glacial fauna was able to exist here at a time that it could not live in Europe. Still the frozen condition of the subsoil, and the formation of ground-ice in the rivers, rendered the preservation of the carcases of mammoths, rhinoceroses, bison, and musk-oxen an easy matter. Passing to south-western Europe, the occurrence of remains of the African elephant in Sicily and Spain, together with the presence of small allied species in the Plistocene of Malta, and likewise of remains of the Barbary sheep and Barbary ape in southern Europe, indicates a free land-communication between Europe and Africa, both by way of the Straits of Gibraltar, and likewise between Italy, Sicily, and Tunis; Malta being then also in 1 Beechey's Voyage to the Pacific and Behring's Straits in H.M.S. "Blossom," Vol. II. L. 22 THE HOLARCTIC REGION. [CHAP. connection with the mainland. Probably also it was by one or both of these routes that the hippopotamus and spotted hyaena passed between Europe and Africa, as it is scarcely likely that the former animal, at least, travelled round by way of Turkey and Syria. Writing of the Maltese islands, Leith-Adams1 observes that "although from their smallness the islands furnish only scant evidences of the complicated and extensive oscillations of level to which the original area has been subjected from first to last, nevertheless the data I have furnished are at the least suggestive, and, in conjunction with the fossilised remains, seem to lead to the belief, that in the first place there was an upheaval of a large tract of land in this portion of the Mediterranean at some period towards or after the close of the Miocene epoch. In the second place, that during the Quaternary [Plistocene] period, the whole, or at least all excepting perhaps the tops of the Benjemma heights and Gozo hills, were again submerged ; and, thirdly, that a re- elevation of the land took place, ending in the present insular fragments. Perhaps in the first case there was a connection or contemporaneity in the upheaval of the Miocene beds of Malta, Sicily, Italy, Candia, the Red Sea, Egypt, Arabia, Cerigo, Azores, Algeria, Southern France, and Spain. Thus the islands of the inland sea may represent portions of a land area now occupied more or less by water. When this area began to sink is not apparent, but the fact that the same elephant and hyaena now living in Africa existed in Sicily, shews that there was a land- connection between the two at a very recent epoch." Regarding the nature of the former connection between Italy, Sicily, and Malta, Dr Wallace2 writes that a comparatively shallow sea or submerged bank incloses Malta and Sicily, and " that on the opposite coast a similar bank stretches out from the coast of Tripoli, leaving a narrow channel, the greatest depth of which is 240 fathoms. Here, therefore, is a broad plateau, which an eleva- tion of about 1,500 feet would convert into a wide extent of land connecting Italy with Africa ; while the same elevation would also connect Morocco with Spain, leaving two extensive lakes to repre- 1 The Nile Valley and Malta (London, 1870), p. 211. 2 Geographical Distribution of Animals, Vol. I. p. 201. IX.] LAND CHANGES. 339 sent what is now the Mediterranean Sea, and affording free com- munication for land animals between Europe and North Africa." Probably the dwarf elephants of Malta were developed from a larger form, closely allied to or identical with the African elephant, after the separation of the island itself from the mainland. With regard to Leith-Adams' idea of the subsequent submergence of Malta, it is pretty certain that this could not have been complete, since that island is inhabited by a large species of weasel (Mustela africand) common to Egypt, and perhaps the south of Italy1; this animal being doubtless a survivor from the old fauna of the Plistocene land connecting Italy, Sicily and Malta with northern Africa. In north-western Europe there are equally conclusive evidences of the connection of the British Islands with the Continent during the Plistocene epoch. On many parts of the English coasts there occur, for instance, submerged forests dating from a comparatively recent epoch, which, when exposed during exceptionally low tides, are seen to contain the stumps of trees in their original upright position,. and with their roots still implanted in the soil. Forests of this kind are found near Torquay and Falmouth, as well as on several parts of the Welsh coasts and in Holyhead harbour ; the submergence which has taken place in the case of the one at Falmouth being estimated at upwards of 70 feet. Again, the pre- glacial Norfolk Forest-bed, so often alluded to in the foregoing pages, affords evidence of an extensive submergence on the east coast of England ; this being supplemented by the Dogger Bank in the North Sea, from which, as already mentioned, such numbers of remains of the mammoth, as well as those of the woolly rhino- ceros and other mammals, have been dredged. Additional evidence in favour of the same subsidence is afforded by the numerous ancient river-channels and valleys found in many parts of Britain, which are situated at depths of from one to two hundred feet below the present level of the land, and frequently cut right across the existing drainage lines, so as to connect valleys now com- pletely distinct. These ancient channels, which are now completely choked with sand, mud, or gravel, have only been revealed by the 1 See Thomas, Proc. Zool. Soc. 1895, pp. 128 — 131. 22—2 340 THE HOLARCTIC REGION. [CHAP. aid of the borer, but their evidence is, nevertheless, unimpeach- able. From these and other lines of evidence, we learn that during the Plistocene epoch not only was England connected with France across the English Channel, but that the land extended up the North Sea at least as far as the Dogger Bank ; the Ouse, the Thames, the Rhine, and perhaps the Elbe originally uniting to form one mighty river, discharging far up in the North Sea. During a portion of this period Ireland was in connection with the British Islands; and it has been suggested by Leith- Adams1 that the connection was with Scotland, owing to the circumstance that, with the exception of the cave-bear, all the living and extinct Irish mammals have been recorded from Scotland, while a number of the English Plistocene mammals appear never to have reached the latter country. On the other hand, Dr R. F. Scharff 2, from a study of the freshwater fishes and molluscs, is of opinion that " Ireland was in later Tertiary times connected with Wales in the south and Scotland in the north ; whilst a freshwater lake occupied the present central area of the Irish Sea. The southern connec- tion broke down at the beginning of the Plistocene period, the northern connection following soon after. There is no evidence of any subsequent land-connection between Great Britain and Ireland." Since the above was written Dr Scharff (Mem. Soc. ZooL France, vol. vin. pp. 436-474, 1895) has further developed his views on the origin of the Irish fauna. He concludes that all the Irish mammals reached the island in the early Plistocene (Forest-bed); such British forms as are unknown in Ireland being considered to have reached Britain later, when Ireland was isolated. Further reference to the former connection or connections between Britain and the Continent will come more conveniently later. The generic and specific mammalian types common to the eastern and western divisions of the Holarctic region Western Division of the having been already referred to, we may at once proceed to the consideration of those characteristic 1 Proc. Roy. Irish Acad. Ser. 2, Vol. III. (1883). 2 Appendix, No. 25. IX.] WESTERN DIVISION. 34! of the western division. And here it may be mentioned in respect to the two areas, that whereas many generic types of animals were unable to pass from the one to the other owing to the high latitude of the strip of connecting land, yet in other cases the geographical limits of the range of certain genera in the Old World form also an important factor in the case. As stated a few para- graphs back, many of the characteristic European Plistocene mammals, such as the hippopotamus, the fallow-deer, and the cave-hyaena, never extended into the Asiatic portion of the Hoi- arctic region, so that these and many other forms never could have had an opportunity of crossing Bering Strait, even had they been capable of existing in such a high latitude. Excluding bats and seals, the following genera of mammals will be found confined to the western half of the Holarctic region, although some of these range southwards into the Sonoran. There are also certain genera which appear to be typically Sonoran, whose range includes part of the western Holarctic region, but these are best considered in the light of intruders from the south l. Among the shrews of the western Holarctic there are two species, viz. Sorex palustris, of the Rocky Mountains, and*S. hydro- dromus, of Unalaska Island, which differ from all their allies in the presence of long fringes of hair to the feet, although they resemble ordinary species of the genus in the characters of their dentition and tail. In consequence of these differences these aquatic shrews have been referred by some writers to a separate genus, under the name of Neosorex ; although such distinction is considered by Dr Merriam unnecessary. There is, however, one genus of Insectivora (Condylura), represented by the star-nosed mole, absolutely charac- teristic of this area. Allied in structure and habits to the Old World moles, which are totally wanting in America, this animal takes its name from the presence of a star-like ring of fleshy appendages at the extremity of the muzzle. The Carnivora include no peculiar genera2; but the Rodentia, 1 It may be well to mention here that the majority of American zoologists regard as genera a number of groups to which the present writer would not be disposed to grant more than sub-generic rank. 2 Mephitis, Taxidea, etc., appear to be of Sonoran origin. 342 THE HOLARCTIC REGION. [CHAP. which, as in the eastern division, are very numerous, comprise one family, as well as several genera, restricted to this area. In the Sciuridce the marmot- like genus Cynomys ranges into the Hoi- arctic, but is considered by Dr Merriam as chiefly characteristic of the Sonoran region; and the same is the case with the white- footed mice (Sitomys) — of which there is but a single Holarctic representative, while the Sonoran species are very numerous — and also with the wood-rats (Neotoma), of which a sub-genus is restricted to the Holarctic. The family peculiar to the region is that of the Haplodontidtz, or sewellels, represented by two species of the genus Haplodon, from the districts west of the Rocky Mountains. Closely allied to the squirrels, these rodents are distinguished from the latter by the absence of postorbital pro- cesses to the frontal bones of the skull, the depressed skull, and the rootless, or hypsodont, cheek-teeth ; all these characters indi- cating a more specialised type. In the Muridce, the voles of the genus Phenacomys connect the more typical members of the group with cricetines like the wood-rats (Neotoma). Several species have been described. A more southern type is the single representative of the allied genus Synaptomys, distinguished by its grooved upper incisors \ its molar teeth resembling those of the lemmings, while its skull is of the same structure as in the true voles. According to Dr Merriam, this animal is restricted to the southern part of the Holarctic area, or what he terms the Transi- tion region. In the same great family the well-known aquatic musk-rat, or musquash (Fiber), may be considered an Holarctic type, since it is found in the " barren-grounds " on the borders of the Arctic sea, although it ranges southwards into the Sonoran. Closely allied to the voles, with which it agrees in the characters of the skull and teeth, this animal differs by the long, compressed, nearly naked, and reticulate tail ; the naked-soled feet being partly webbed, and the whole body adapted to an aquatic mode of life. Its fossil remains occur in the Plistocene of the United States. It is noteworthy, as a negative characteristic of the Holarctic area, that no members of the exclusively New World family Geomyidce are found within its limits. On the other hand, in the family Dipodida, the j urn ping-mice of the genus Zapus, of which several species are recognised by North American zoologists, are solely IX.] WESTERN DIVISION. 343 Holarctic, the typical Z. hudsonianus dating from the Plistocene epoch. A distinctive feature of the western Holarctic region is the absence of true porcupines (Hystrix), their place being taken by the Canadian porcupine (Erethizon], which belongs to the same sub-family as the South American porcupines, although distin- guished, among other characters, by its short and non-prehensile tail. It is a native of the wooded portions of Canada and the United States, and its remains have been discovered in a cave in Pennsylvania. FIG. 69. ROCKY MOUNTAIN GOAT (Haploceros montamis] Among the ungulates, the remarkable animal known as the Rocky Mountain goat, which alone represents the genus Haplo- ceros, and differs from all other ruminants by the extreme shortness of the cannon-bone in both the front and hind limbs, is exclusively an inhabitant of the western Holarctic region. The same is now 344 THE HOLARCTIC REGION. [CHAP. the case with the musk-ox (Ovibos), but as this animal ranged over Europe and northern Asia during the Plistocene, it can scarcely be regarded as distinctive of the western area. Of other peculiar New World ungulates, the prong-buck (Antilocapra) and certain deer of the genus Cariacus are found within the Holarctic region, but the former seems to be essentially a Sonoran type, while the latter, although probably also of Sonoran origin, occurs through Central and South America. The Tertiary genera of mammals peculiar to North America may be best considered in the chapter devoted to the Sonoran region, to which they for the most part belong ; and this portion of the subject may be accordingly concluded by tabulating the existing genera or groups peculiar to the area under consideration. These will stand as follows, viz. : — Insectivora. SORICID.E. Sorex. The sub-genus or genus Neosorex. TALPID^E. Condylura. Rodentia. HAPLODONTIDJE. Haplodon. MURID^:. Phenacomys. Synaptomys. Confined to the southern portion of the area. Fiber. Enters Sonoran. DIPODID^E. Zapus. HYSTRICID.E. Erethizon. Ungulata. BOVID^E. Haploceros. Even if we add to the above certain other sub-generic types, such as the spruce-squirrels (Tamiasdurus) and the bushy-tailed wood-rats (Teonoma), and likewise take into account the number of Old World types (in many cases widely-distributed ones) that are absent, it can scarcely be urged that such an assemblage is sufficient to constitute a zoological region by itself. Those of my readers desirous of consulting lists of the species inhabiting the Arctic and Boreal zones of Dr Merriam, will find them in his memoir1. 1 Appendix, No. 19, pp. 24, 25. IX.] WESTERN DIVISION. 345 In America, probably owing to the north and south trend of the mountain-ranges, the glacial period has had an even more marked effect than in the Old World. On this subject Dr Merriam1 writes that "not only are the pre-Plistocene animals and plants now represented imperfectly and in greatly reduced num- bers, but the areas at present inhabited by their descendants, except in the case of the Boreal forms, are insignificant in com- parison with their former extent. It should be remembered that the refrigeration of the glacial epoch has only in part disappeared. In earlier Pliocene times, characteristic representatives of sub- tropical faunas and floras existed northwards over much of the United States and Canada, and in still earlier times reached the Arctic circle. During the advance of cold in the glacial epoch these forms were either exterminated or driven southward into the narrow tropical parts of Mexico and Central America. The retreat of cold at the termination of this period was not complete, and our continent has never regained its former warmth. Hence the expelled species were not permitted to advance more than a short distance into the region formerly occupied by them, and the tropical species have been held back, and at the present day are not found except along the extreme southern confines of our territory [the United States]. For example, peccaries in early Plistocene times ranged northward over a large part of western North America, while at present they are restricted to parts of Texas and Louisiana below the Red River of the south ; and capivaras, tapirs and other tropical forms whose fossil remains have been found in many parts of the United States have not been able to return. The same is true of plants, for the palms, tree- ferns, and numerous other tropical types that formerly ranged over much of our country are now either altogether extinct or exist only in the tropics. "The llama and many plants now inhabiting the Andes may be looked upon as representing a class of cases in which Boreal forms were driven so far south that they actually reached the great mountain-system of South America and spread southward over its elevated plateaus and declivities to the extreme end of the conti- nent in Patagonia and Tierra del Fuego." 1 Appendix, No. 19, p. 44. 346 THE HOLARCTIC REGION. [CHAP. Coming to the consideration of sub-regions, we have first of all the Arctic sub-region, which corresponds to the Boreal sub-region of Dr Heilprin, and the Arctic zone of the Boreal region of Dr Merriam, and is of circumpolar extent. According to the former writer, in the Old World it may be defined as the tract lying to the north of a line starting from about the 66th parallel of latitude on the Norwegian coast, and passing south-eastwards to the coast of eastern Asia in about the 5oth parallel, thus including the greater part of Kamschatka, and Amurland. In America, according to Dr Merriam's map, after running just inside the shores of Newfound- FIG. 70. MUSK-OX (Ovibos moschatus). land and Labrador, the boundary line bends southwards after passing Cape Chudleigh to coincide with the southern shore of Hudson Bay, and then takes a north-westerly direction so as to include within the sub-region only a narrow strip on the north- eastern coast of Alaska, and a somewhat broader one on the north-western shore of the same. In the Old World the boundary line coincides approximately with the northern limit of the cultiva- tion of cereals, and also with that of the southern migrations of the reindeer; but in America certain reindeer (which are regarded IX.] ARCTIC SUB-REGION. 347 by the American zoologists as specifically distinct from thecircum- polar "barren-ground" variety) extend considerably further to the south. For the most part of its extent, the mammals inhabiting this sub-region are few in number, a large proportion of them having a circumpolar range. Among them may be included the Arctic fox, polar bear, wolverene, the ermine or stoat, the eastern and western species of lemming (Myodes), the banded lemming (Cuniculus torquatus), the Arctic vole (Microtus nitilus}, Parry's suslik (Spermophilus empetra], the musk-ox, and the reindeer ; several of these being restricted to the sub-region. The sea-otter (Latax) frequents the shores of Alaska and Kamschatka, but also ranges as far south as the Kurile Islands and California, so that it is not confined to the sub-region. During the Plistocene epoch, as we have seen, such animals as the mammoth, horse, bison, and tiger were inhabitants of this tract ; the latter animal being still found in eastern Siberia. Towards Amurland and the Kams- chatkan peninsula, the fauna becomes somewhat less scanty ; the large Kamschatkan sheep (Ovis nivicola) being here met with, as well as a true deer, and the brown bear. Of other groups of animals inhabiting the more typical portions of this region, there may be noticed among the birds the ptarmigan (Lagopus], the snowy owl (Nyctea scandiaca), the Greenland falcon (Falco candicans), the eider-duck (Somatcria mollissima), as well as various species of divers (Colymbus] and guillemots (Uria and Lomvia}, together with the little auk (Mergulus alle\ Dr Heilprin writes that " Captain Markham observed the footpiints of the polar hare in the snow-bound ice in latitude 83° 10', and the antlers of a reindeer were picked up by the officers under Sir George Nares, in latitude 82° 45' (Grinnell Land). A skeleton of the latter animal, recently picked by wolves, was also obtained in latitude 80° 27'. Traces of the rock-ptarmigan (Lagopus rupestris] have been met with as far north as latitude 83° 6', and the snow- bunting (Pkctrophanes nivalis) in latitude 82° 33'. The reptile- fauna is very limited, no serpent, apparently, passing beyond the sixty-seventh parallel of latitude, and no lizard above the seventieth. The fishes, which include the common perch and pike, are mainly salmonoids. Insects are fairly numerous, and even in the far north the number of species is considerable." 348 THE HOLARCTIC REGION. [CHAP. During the Plistocene the region within the Arctic circle en- joyed a decidedly less rigorous climate than it at present pos- sesses. In Baron von Toll's recent expedition to the New Siberian Islands1, where, as previously stated, remains of the tiger have been obtained, it was discovered " that under the perpetual ice, in a freshwater deposit, which contained pieces of willow and bones of post-tertiary mammals (the mammoth-layer) were complete trees of Alnus fruticosa, fifteen feet long, with leaves and fruit. It was thus evident that during the mammoth-period tree-vegetation reached the seventy-fourth degree of latitude, and that its northern limit was at least three degrees further north than it is now." The next sub-region is the European, which may be taken to include all that part of Europe lying between the Arctic sub-region in the north, and the line of the Pyrenees and Alps, continuing eastwards along the northern shore of the Black Sea to the Caucasus and the Caspian Steppes. This area includes the typical fauna of the eastern Holarctic region, among its more or less characteristic mammals being (in the north) the elk — also ranging into America — , the red deer (unknown in America, but represented by a variety in North Africa), the roe, the bison, the chamois, the Alpine ibex, the typical variety of the brown bear, the badger, the wolverene (in the north), the Alpine marmot (Arctomys marmotta), the dormouse, hamster, mole, and hedgehog \ several of these being, however, common to the Arctic and Central Asian sub-regions. The des- mans (Myogale) are restricted to this sub-region ; and the same was probably the case with the aurochs (Bos taurus, var. primi- genius], the ancestral stock of our domestic cattle. Finally, the Caucasus is the home of two or three peculiar species of goats (Capra cylindricornis and C. caucasica] known as ture. It will be unnecessary, even if this could be accomplished, to give a complete list of the mammalian fauna of this sub-region, but it is essential to refer to the comparative poverty of the fauna of the British Islands as compared with that of the Continent. The following list includes all the mammals (exclusive of bats) known to have inhabited the British Islands within the historic 1 See Knowledge, 1895, p. 106. IX.] EUROPEAN SUB-REGION. 349 period ; those which are now extinct having an asterisk prefixed to them, while such as occur in Ireland have the letter I added. Those that have been introduced by man have a t before them. The list stands as follows, viz. : — Hedgehog. Erinaceus europaeus. I. Mole. Talpa europsea. Common Shrew. Sorex araneus. Lesser ,, ,, minutus. I. Water-Shrew. Crossopus fodiens. Wild Cat. Felis catus. *Wolf. Canis lupus. I. Fox. „ vulpes. I. Pine-Marten. Mustela martes. I. Polecat. ,, putorius. Stoat. „ erminea. I. Assogue. ,, hibernica. I. Weasel. ,, vulgaris. Badger. Meles taxus. I. Otter. Lutra vulgaris. I. ^Brown bear. Ursus arctus. I. Squirrel. Sciurus vulgaris. I (? introduced), * Beaver. Castor fiber. Dormouse. Muscardinus avellanarius. Harvest-Mouse. Mus minutus. Wood-Mouse. ,, sylvaticus. I. Yellow-necked Mouse. ,, flavicollis. Common Mouse. ,, musculus. I. f Black Rat. „ rattus. I. t Brown Rat. ,, decumanus. I.. Common Field:Vole. Microtus agrestis. Bank-Vole. „ glareolus. Water-Vole „ amphibius. Common Hare. Lepus europaeus. Mountain Hare. ,, timidus. I. t Rabbit. ,, cuniculus. I. *? Wild Cattle. Bos taurus. Red Deer. Cervus elaphus. I. 350 THE HOLARCTIC REGION. [CHAP. t Fallow Deer. Cervus dama. I. Roe Deer. Capreolus caprea. *Wild Boar. Sus scrofa. I. The total number in this list is only 28, out of which at least four are introduced. With the exception of the recently-described assogue1, which is intermediate between the stoat and the weasel, and is peculiar to Ireland, the whole of these mammals are com- mon to the Continent. As shown in an earlier portion of the present chapter, during the Plistocene epoch Britain possessed a fauna apparently identical with that of the Continent ; and there must accordingly be some good reason for its present poverty in mammalian life as compared to the latter area. The difference is accounted for by Dr Wallace, through the occurrence of one or more periods of subsidence, which took place during the close of, or subsequent to, the Glacial epoch ; after which England again became united to the Continent, when its present fauna entered, the period of connection being, however, of comparatively short duration, and thus permitting of the passage of only a moiety of the continental forms, or those which happened at the time in question to be inhabiting the districts nearest to the connecting line. Only a certain number of the mammals which thus crossed into Britain have ever succeeded in reaching Scotland ; and it is from this country, if we accept the views of Dr Scharff, referred to above, that Ireland appears to have received its still more im- poverished mammalian fauna. It will be seen that the foregoing hypothesis attributes the clean sweep supposed to have been made of the original British fauna to the effects of submergence, and not to the ice-sheet. On the other hand, Mr G. W. Bulman 2, who doubts whether the sub- mergence has been sufficient for this, attributes such extermination as he believes to have taken place solely to the effects of an ice- sheet. And he further believes that a number of the original British mammals survived in the southern and south-western counties of England, whence they re-populated Britain on the disappearance of the ice-sheet, without there having been any 1 See Thomas, Natural Science, Vol. vi. p. 377 (1895). - Appendix, No. 12. IX.] CENTRAL ASIAN SUB-REGION. 351 subsequent connection with the Continent. The difficulty con- nected with this explanation is that it apparently necessitates a pre-glacial or early glacial age for the mammaliferous deposits of the Thames valley, which are almost certainly inter-glacial or post-glacial. The whole subject of glaciation is, however, so complicated and involved, that it is almost impossible to form workable theories as to the exact mode of the repopulation of Britain after the changes which took place during the glacial epoch. In contrast to the British Isles, which are eminently of the continental type, may be cited Iceland, lying near the border-line between the Arctic and European sub-regions, which is as markedly oceanic in its character. Beyond an occasional ice- borne polar bear, Iceland possesses only the Arctic fox, and a mouse, which has been stated to be a peculiar species ; the fox having doubtless been originally introduced from the north on floating ice. According to the scheme of Dr Heilprin, the next sub-region on the list is that of Central Asia, which includes the countries bounded on the west by the European, sSb-regi™*1*11 and on the north by the Arctic sub-region, and extends eastwards as far as Mantchuria and China proper, being bordered on the south by the Kuenlun and Nanshan mountains1. A large portion of the western districts of this tract are open steppes or deserts ; and in such tracts several peculiar types of rodents, such as the Kirghiz jerboa (Alactaga) and the Yarkand jerboa (Euchoretes\ are met with, while the saiga antelope (Saiga), and the Mongolian gazelle ( Gazella guttu rosd) are likewise charac- teristic types. Susliks (Spermophilus), marmots (Arctomys), and picas are very abundant ; and the place of the European wild cat is occupied by Pallas's cat (Felis manul), the tiger being also sparingly found in the western districts, where the ounce is like- wise met with. In part of this sub-region the red deer is replaced by a variety or species known as Cervus xanthopygus, while Yarkand is the home of a variety of the Kashmir stag (C. cash- mirianus], and the Thian-Shan possesses the very fine and wapiti- 1 Dr Heilprin included the Tibetan plateau in this sub-region. 352 THE HOLARCTIC REGION. [CHAP. like form described under the name of C. eustephanus ; all these deer being mostly inhabitants of forest-districts. The Tatarian roe (Capreolus pygargus), inhabiting suitable localities in the moun- tains forming the watershed between the Russian and Chinese empires and Turkestan, is also generally regarded as specifically distinct from its western ally. The sub- region is also the chief home of the magnificent sheep known as argali, among which the splendid Pamir sheep (Ovis poll) ranges from the Pamirs to the Altai, while the true argali (O. ammon) — if the Tibetan O. hodg- soni be really distinct — is also restricted to this sub-region, where it is now confined to northern Mongolia, although it formerly inhabited the Altai. The ibex of the Altai is, however, identical with the Himalayan and Tibetan Capra sibirica. Although included by Dr Heilprin in the preceding, the area typified by the Tibetan plateau is regarded by Dr Blanford ' as constituting a sub-region by itself. Typically this region is bounded on the north by the ranges of the Kuenlun, Altyn Tag, and Nanshan, and extends eastwards to China proper, while to the west it must be taken to include Ladak and the upper Indus valley as far as Gilgit2. To the south it extends to the main chain of the Himalaya. The following list of mammals is given by Dr Blanford as distinctive of this sub-region ; the names of such species and genera as are entirely or mainly confined to the area being printed in italics. Insect! vora. Crocidura aranea. Nectogale elegans. Carnivora. Paradoxurus laniger. Canis lupus, var. laniger. ,, vulpes, var. flavescens. ,, ferrilatus. ,, deccanensis, var. Mustek foina, var. ,, larva ta, 1 Proc. Zool. Soc. 1893, p. 449. 2 See Blanford, Fauna of British India, Mammalia, p. v. IX.] TIBETAN SUB-REGION. 353 Carnivora (cont.\ Mustela canignla. „ alpina, var. temon. „ erminea. Meles leucura. „ albogularis. sEluropus melanoleuctts. Ursus pruinosus. Rodentia. Eupetaurus rinereus. Arctomys himalayanus. „ robustus. Mus sublimis. Microtus blythi. ,, strauchi. „ przevalskii. Siphneus fontanieri. Lagomys curzonice. ,, rutilus. ,, erythrotis. „ melanostomus . ,, ladacensis. Lepus oiostolus. ,, hypsibius. Ungulata. Equus hemionus, var. kiang. Bos grunniens. Ovis hodgsoni. „ vignei, var. ,, nahura. Capra sibirica. Pantholops hodgsoni, Budorcas taxicolor. Gazella picticaudata. Cervus affinis. ,, thoroldi. Moschus moschiferus. L. 23 354 THE HOLARCTIC REGION. [CHAP. This list includes all the species inhabiting the plateau at elevations exceeding 12,000 feet. Dr Blanford writes that "many of the forms named only inhabit small portions of the area, and whilst Bos grunniens, Pantholops hodgsoni, and Gazella picti- caudata, with several rodents, appear to be peculiar to the high plateaus above 14,000 feet, the two species of Cervus are pro- bably found in brushwood at a rather lower elevation in the more broken region of Eastern Tibet, where the rainfall is heavier and the vegetation more abundant. "As was printed in the paper in the Geological Magazine1, there is, so far as I am aware, no equally peculiar mammalian fauna to be found in any continental area of equal extent, and for a parallel it is necessary to turn to some island like Celebes, that has long been isolated from all surrounding lands." This, however, is not all, for there occur at Hundes, on the Tibetan plateau, mammaliferous strata yielding, among other remains, bones of a rhinoceros, and of an antelope which is apparently generically identical with the chiru (Pantholops), now inhabiting the same area. The isolation and development of this most peculiar fauna is intimately connected with the date of elevation of the Himalaya. After pointing out that both the fossil chiru and the fossil rhinoceros appear to have inhabited the area when it had attained something approaching its present enormous elevation, Dr Blanford2 writes as follows : " Bearing in mind that the isolation of the Tibetan plateau is far less perfect as regards mammals than that of any island, and that some of the forms — the Carnivora especially — found in Tibet are evidently very recent immigrants, it is a reasonable conclusion that the peculiar fauna of the Tibetan plateau has been distinct from that of neighbouring countries since middle Tertiary times. " But what has caused the isolation of the Tibetan fauna ? Why in this one continental tract is there a generic and specific differentiation of the mammalia, of which no other example exists ? There is only one character in which Tibet is different from other continental areas, its great height. This alone renders the climate of Tibet so different from that of other parts of 1 Decade 3, Vol. ix. p. 161 (1892). 2 Geol. Mag. op. cit. p. 165. IX.] MANTCHURIAN SUB-REGION. 355 Central Asia, which are equally cold and barren. It seems a reasonable inference that the elevation of the Tibetan plateau dates back to middle Tertiary times. " It is of course probable that the elevation was gradual ; and although the area may have been sufficiently high at the close of the Miocene period to produce a difference in climatal conditions, the greater part of the upward movement may have been post- Miocene, and a great part post-Pliocene." Bordering as it does upon the tropics, where it abuts against the Oriental region, the Mantchurian sub-region is not easy to define, since the intermingling of Holarctic and Oriental types is very strongly marked on its southern confines. Starting somewhere about the Amur river, it may, however, be taken to include the Japanese islands, Mant- churia, Corea, and northern China ; its southern limit being placed approximately in the latitude of Fuchau. Westwards it may be taken to include Moupin, in Eastern Tibet, although this district is referred by Dr Wallace to the Oriental region. From all the other sub-regions, with the exception of the Mediterranean, the Mantchurian is distinguished by the presence of monkeys belonging to the genera Macacus and Semnopithecus, some of these occurring in Japan and others in Eastern Tibet. Of the latter, one (Semnopithecus roxellance) is peculiar, and the other is identified by Mr H. O. Forbes with the widely-spread Oriental Macacus arctoides. Among the Carnivora, the Oriental genus Helictis enters this sub-region, one species occurring in the neighbourhood of Shanghai ; while Japan is the home of a peculiar long-haired dog ( Cants procyonides), which is frequently separated generically under the name of Nyctereutes, although it unquestion- ably pertains to the typical genus. Perhaps, however, the most characteristic mammals are the deer. Foremost among these are a group of small deer belonging to the genus Cenms, and distin- guished from the red deer group by the invariable absence of a bez-tine to the antlers, each of which has but four points. These deer are further characterised by the coat of the adult being spotted in summer with white, but uniformly brown in winter, and also by the black lateral margins to the white blaze on the hind- quarters. The species include the Japanese deer (C. sica), 23—2 356 THE HOLARCTIC REGION. [CHAP. common to Japan and North China, the larger Mantchurian deer (C. mantchuricus\ and Dybowski's deer (C. dybowskii) from the upper Ussuri district of Mantchuria, in the neighbourhood of Vladivostock. Elsewhere the group is represented in Formosa, and also in the Caspian provinces of Persia. In addition to these, there are the hornless Chinese water-deer (Hydropoles), and the two species of tufted deer (Elaphodus)\ the latter being closely allied to the Oriental muntjacs. What is known of the palseonto- logical history of the southern portion of this area indicates that during the Pliocene epoch its mammalian fauna was closely allied to that of the Siwalik Hills, thus showing that at this time there was no distinction between the Oriental and Holarctic regions, which even now grade imperceptibly into one another in this district. The remains of fossil elephants from Japan1 are referable to Elephas clifti, insignis, and namadicus, of which the two first are common to the Siwaliks, while the third occurs typically in the Plistocene Narbada beds of India. From the known distribution of these elephants, it is probable that Japan was connected with the mainland during the Pliocene epoch by way of the Corean peninsula, although Dr Wallace is of opinion that its latest con- nection was to the north. Of existing animals common to Japan and the mainland, allusion has already been made to Cervus sica : and another common type is the giant salamander Megalobatra chus. The latter genus is represented in a fossil state in the Miocene of Baden, and as it is closely allied to the North American CryptobranchuS) there is clear evidence of the eastern migration of this ancient type, of which the two survivors are respectively con- fined to China and Japan on the one hand, and North America on the other. Further evidences of affinity between the fauna of Japan and North America are afforded by the circumstance that one species of the mole-like genus Urotrichus is confined to the former islands, while the other is an inhabitant of the north- western districts of the latter continent. The sea-otter (Latax] is likewise common to the coasts of Japan, the Kurile Islands, and Kamschatka, and the Pacific shores of North America. More remarkable, however, is the fact that a North American scincoid 1 See Naumann, Palceontographica, Vol. xxvm. Art. r (1881). IX.] MEDITERRANEAN SUB-REGION. 357 lizard (Eumeces quinquelineatus] is represented in Japan by a form (E, marginatus] so closely allied that the two were long considered inseparable, although they are now regarded as distinct1. All these facts are indicative that Japan was formerly joined to both Corea and Kamschatka, whence land was continued across Bering Strait to unite the Old World with Alaska. Although, as already stated, the Mediterranean or Tyrrhenian sub-region has strong claims to be regarded as Mediterra representing a region by itself, it may be more con- nean Sub- veniently considered here than later on in the chapter. In addition to such parts of Africa and Arabia as lie to the north of the Ethiopian region, this sub-region includes Spain, those parts of Europe situated south of the Alps, together with Turkey, Asia Minor, Persia, Baluchistan, and Afghanistan. Whether Kashmir should be regarded as an aberrant outlier of this region, I am not yet satisfied. Although gerbils (Gerbillus*) are also found in the Oriental and Ethiopian regions, their distribution in the Holarctic is very nearly coincident with the limits of the present sub-region. Whereas to the north of the Mediterranean Sea a large pro- portion of the mammals are more or less typically Holarctic, in North Africa and Syria those with an Ethiopian facies are met with, and an Oriental element makes its appearance in the eastern districts of the sub-region. Even in Africa, however, some of the forms have an Oriental facies, the Barbary ape (Macacus inuus) belonging to a genus whose home is now in the Oriental region, and which is totally unknown in the Ethiopian. As a wanderer from the Ethiopian region, mention may first be made of a species of jumping -shrew (Macroscelides) met with in Barbary, while among the octodont family of the rodents, the gundi, forming the sole representative of the genus Ctenodactylus, has its nearest allies in Ethiopia, although it is confined to North Africa. The Barbary ape, although occurring on the rock of Gibraltar, where it may have been introduced, is otherwise confined to North Africa. 1 See Boulenger, Cat. Lizards, Brit. Mtis. Vol. III. p. 369. 2 Many writers separate certain species as Meriones, but as the two groups are connected by G. indicus (see Lataste, Proc. ZooL Soc. 1884, p. 88), such distinction seems superfluous. 358 THE HOLARCTIC REGION. [CHAP. In the Carnivora, the striped hyaena, which is also an inhabitant of India, is widely spread in this sub-region, ranging through western Asia to northern Africa. The common genet (Genetta vulgaris\ which belongs to a genus otherwise exclusively Ethio- pian, is mainly confined to this region, inhabiting southern France, Spain, Turkey, North Africa, and Palestine. A nearly similar distribution characterises the common mungoose, or ichneumon (Herpestes ichneumon], which frequents southern Spain, Asia Minor, North Africa, and Palestine. The large weasel (Mustela africanus) common to Egypt, Malta, and perhaps the south of Italy has been already referred to in an earlier part of this chapter. In addition to Ctenodactylus, the rodents possess another and more widely-spread generic type confined to this sub-region in the form of the great mole-rat (Spalax typhlus), whose range includes south- eastern Europe, Persia, Mesopotamia, Syria, and Egypt. In the same order the common porcupine (ffystrix cristata], although ranging into West Africa, is found but little, if at all, to the north of the present sub-region, where it is common to northern Africa and southern Europe. Among the ungulates the addax antelope (Addax nasomacu- lata\ although allied to Ethiopian types, is solely Mediterranean, its home being North Africa and Syria. More closely allied to the Ethiopian fauna are certain hartebeests of the genus Bubalis, the smaller of which (B. mauritanica] is common to North Africa, Syria, and Arabia, while the second {B, major] inhabits Tunis. The same is the case with the Beatrix antelope (Oryx beatrix] of Western Arabia and Bushire. In gazelles, this sub-region is remarkably rich, doubtless from the number of sandy or desert tracts it contains. Algeria is the habitat of the three species known as Gazella loderi, G. kevella, and G. rufina, while G. dorcas ranges through Egypt, Algeria, Syria, Palestine, and a part of Asia Minor, and G. subgutturosa roams from Persia through Afghanistan and Turkestan. The aberrant sheep known as the arui (Ovis tragelaphus] is now restricted to North Africa; and the mouflon (O. musimon), although its fossil remains have been found on the Continent, appears to be now restricted to Corsica. An- other species peculiar to the sub-region is the Armenian sheep (O. gmelini) of eastern Persia and Asia Minor, represented by a IX.] KASHMIR. 359 closely-allied form in Cyprus. Of the goats, the Spanish ibex (Capra pyrenaica) is restricted to the mountains of Spain; while the Sinaitic ibex (C. sinaitica) represents the group in Palestine and upper Egypt. Among the Cervidcz, the two species of fallow- deer were originally confined to this area, the common Cervus dama being a native of the Mediterranean countries, while the Persian C. mesopotamicus is found in the mountains of Luristan, in Mesopotamian Persia. In North Africa the ordinary red deer is represented by a variety distinguished by the invariable absence of a bez-tine to the antlers. A connection with the Tibetan sub- region is afforded by the wild asses inhabiting the desert-plains between the Red Sea and the Indus, since both these and the Tibetan form are but varieties of a single species (Equus hemi- onus). Lastly, Ethiopian affinities are exhibited by the occurrence of a species of hyrax (Procavid] in Syria. In the early part of the present century the hippopotamus still inhabited lower Egypt, while, as we have seen, the lion, which is now common in parts of Persia, was found within the historic period in Thrace. At a still earlier date, both these animals, as well as the spotted hyaena, extended as far north as England. On the whole, therefore, the fauna of this sub-region is a very mixed one ; and this fact, together with the difficulty in defining its boundaries, suggests the need of further deliberation before the area is raised to the rank of a separate region. The former con- nections between southern Europe and Africa having been alluded to in an earlier part of the chapter, require no further notice in this place. Very difficult to determine is the position which should be given to the valley of Kashmir, since its fauna exhibits such a mingling of Oriental and Holarctic types that it might almost be as well assigned to one region as the other. Holarctic affinities are, however, exhibited by the occur- rence of a species of the red deer group, Cervus cashmirianus, and likewise by one of the rodent genus Sminthus, of which the second species inhabits northern and eastern Europe, and the third Kansu, in western China. A variety of the brown bear is also indicative of Holarctic affinities, and this is still more markedly the case with the spiral-horned goat known as the 360 THE HOLARCTIC REGION. [CHAP. markhor (Capra falconeri], of which one variety inhabits the Pir Panjal range, on the south side of the valley, while the others are found in the districts to the north and west of Kashmir. The musk-deer, again, is another essentially Holarctic type. On the other hand, the occurrence of a langur -(Semnopithecus) and a ma- caque (Macacus) points to a connection with the Oriental fauna ; and a Kashmir mungoose (Herpestes auropunctatus] is identical with one from India. There are, however, none of the exclusively Oriental genera in Kashmir ; and this fact, coupled with the absence of all deer of the sambar-group, leaves little doubt that the valley really belongs to the Holarctic. Whether it should be regarded as pertaining to the Mediterranean sub-region, or as forming a distinct sub-region by itself, must be reserved for future consideration. Passing to the western division of the Holarctic region, the tract lying to the southward of the circumpolar Arctic sub-region, designated by Dr Merriam the Boreal zone of his Boreal region, may be conveniently termed the Canadian sub-region. Its northern boundary is, of course, identical with the southern limits of the Arctic sub-region, and the area includes the greater part of the Dominion of Canada, although a long strip runs down the line of the Rocky Mountains, and another along the Pacific coast, into the United States. Indeed, Dr Merriam includes in this sub-region all the higher plateaus of Wyoming and Colorado, so that the sub-region embraces a number of small disconnected areas on its south-western ex- tremity, and it is consequently impossible to define its limits by description. It may be stated, however, that on the eastern side of the continent the sub-region extends from Hudson Bay to the middle of Lake Michigan, while on the western coast it stretches from near the extremity of Alaska to San Francisco ; a big loop extending northwards of Montana nearly to latitude 55°. The mammalian fauna of the Canadian sub-region is that of the western division of the Holarctic region generally, and includes those forms mentioned on page 344. According to Dr Merriam, the following genera from this sub-region do not range further south than the undermentioned Transition zone ; namely, Condy- lura, Urotrichus, Gulo, Latax, Arctomys, Haplodon, Phenacomys, IX.] TRANSITION ZONE. 361 Myodes, Cuniculus, Zapus, Erethizon, Lagomys, Cervus, Alces, Rangifer, and Haploceros. On the other hand, the following, which are as clearly of northern origin, penetrate as far south as the Sonoran region, which some of them enter. These are Sorex, Mustela (only the members of the sub-genus Putorius), Ursus, Fiber, Microtus, Castor, Tamias, Bos, and Ovis. Between the Canadian sub-region of the Holarctic and the Sonoran region is interposed a tract whose fauna contains a mixture of Canadian and Sonoran forms, Transition Zone. and it is consequently termed by Dr Merriam the Transition zone. Under this somewhat indefinite title the area may best be left. It is described by the author just cited as follows1: "The humid division of this zone, known as the Alleghanian fauna, covers the greater part of New England (except Maine and the mountains of Vermont and New Hampshire), and extends westerly over the greater part of New York, southern Ontario, and Pennsylvania, and sends an arm south along the Alleghanies, all the way across the Virginias, Carolinas, and eastern Tennessee, to northern Georgia and Alabama. In the Great Lake region this zone continues westerly across southern Michigan and Wisconsin, and then curves northward over the prairie-region of Minnesota, covering the greater parts of North Dakota, Manitoba, and the plains of the Saskatchewan ; thence bending abruptly south, it crosses eastern Montana and Wyoming, including parts of western South Dakota, and Nebraska, and forms a belt along the eastern base of the Rocky Mountains in Colorado and northern New Mexico, here as elsewhere occupying the interval between the Upper Sonoran and Canadian zones. " In Wyoming the Transition zone passes broadly over the well-known low divide of the Rocky Mountains, which affords the route of the Union Pacific railway, and is directly continuous with the same zone in parts of Colorado, Uta, and Idaho, skirting the Canadian boundaries of the Great Basin all the way around the plains of the Columbia, sending an arm northward over the dry interior of British Columbia, descending along the eastern base of the Cascade Range and the High Sierra to the southern extremity 1 Appendix, No. 19, p. 30. In this extract the word Canadian has been substituted for Boreal. 362 THE HOLARCTIC REGION. [CHAP. IX. of the latter, and occupying the summits of the Coast Ranges in California and of many of the desert ranges of the Great Basin. "The Transition zone, as its name indicates, is a zone of overlapping of Canadian and Sonoran types. Many Canadian genera and species here reach the southern limits of their distribu- tion, and many Sonoran genera and species their northern limits. But a single mammalian genus (Synaptomys) is restricted to the Transition zone, and future research may show that it inhabits the Canadian region also." FIG. 71. MUSQUASH (Fiber zibet hi cus). The writer adds, however, that there are a considerable number of species — mostly rodents — restricted to this zone. The follow- ing Canadian genera, namely, Condylura, Urotrichus, Ursus, Arctomys, Tamias, Fiber1, Zapus, Erethizon, Cervus, and Ovis, almost or completely disappear in this zone ; while the following intruders from the Sonoran, namely Scalops, Bassariscus, Spilo- gale, Perognathus, Thomomys, Geoitiys, Cynomys and Antilocapra do not range further north, several of them, indeed, only intruding into the zone in a small area in the west. 1 Penetrates the Sonoran along the lines of streams where cool currents of air are carried down. CHAPTER X. THE SONORAN REGION. Limits — Characteristics of Mammalian Fauna — Extinct Groups of Mammals characteristic of Western Arctogaea — Distinctness of the Region — Dual Origin of Groups. As stated in the introductory chapter, wherever one zoological region of the globe has no definite physical barrier by which it is separated from the next well-marked region, there must always occur an intermediate tract where the characteristic types of the faunas of the two regions inosculate and intermingle. That this is the case with that area of North America denominated the Sonoran region has been indicated at the close of the preceding chapter, and the existence of the Transition zone, which seems, on the whole, to pertain to the Holarctic region, unfortunately prevents the Sonoran from being defined with the precision which would be possible had this area a high mountain-barrier on its northern frontier. In a map of the small dimensions of the one accompanying this volume it is impossible to show with any Limits. attempt at accuracy the complex nature of the northern boundary of this region, which will, however, be found accurately laid down in the map illustrating Dr Merriam's memoir1. According to the latter writer, " the Sonoran region as a whole stretches across the continent from the Atlantic to the Pacific, covering nearly the whole country south of latitude 43°, and reaching northward on the Great Plains and Great Basin to about latitude 48°. It is invaded from the north by three principal intrusions of Canadian2 forms along the three great mountain- 1 Appendix, No. 19. 2 Boreal in the original. 364 THE SONORAN REGION. [CHAP. systems already mentioned [Alleghanies, Rocky Mountains, and Cascade and Sierra Nevada ranges] ; while to the southward it occupies the great interior basin of Mexico, and extends into the tropics along the highlands of the interior. It covers also the peninsula of lower California, the southern part of which seems entitled to rank as an independent subdivision." Later it is stated that the region "may be divided by tempe- rature into two principal transcontinental zones, Upper Sonoran and Lower Sonoran ; and each of these in turn may be subdivided into arid and humid divisions." The proposal to form a separate region for such an insignifi- cant area as the southern extremity of California seems unnecessary, although its fauna may differ considerably from that of the typical Sonoran. Omitting mention of the bats, the mammalian genera charac- Characteris teristic of the Sonoran region may now be taken into tics of Mam- consideration. Commencing with the Insectivora, tnalian Fauna. , , .. the Sonadcz are represented by the peculiar genus Notiosorex, which is closely allied to the Oriental Soriculus, but has only 28 in place of 30 teeth. Of this genus the two species do not range north of this region, although they also enter Central America1. The short -tailed, or earless shrews (Blarina), with either 32 or 30 teeth, are also mainly Sonoran, although ranging northwards into the Holarctic, and southwards into Guatemala. In the Talpida the three species of the mole-like genus Scalops, characterised by having 36 teeth, webbed hind feet, and a short and nearly naked tail, are mainly Sonoran, although passing into the Transition zone. On the other hand, the two species of Scapanus, distinguished by the possession of 40 teeth, and the hairy tail, have a distribution very similar to Blarina, although they do not enter Central America. In the Carnivora the raccoon-family (Procyonidce) is very strongly represented, although none of the genera are absolutely peculiar to the region. The genus Bassariscus — a near ally of the true raccoons, and possessing two species — is nevertheless mainly Sonoran, although it ranges into the Transition zone of the 1 Teste Dobson. X.] ITS FAUNA. 365 Holarctic and also into Central America. The true raccoons, on the other hand, cannot be regarded as distinctive of the region, since they range from South America into the Canadian sub- region of the Holarctic ; and the coatis (Nasua) are now highly characteristic of the Neogaeic realm. Indeed Dr Merriam con- siders both genera as intruders from the latter realm, but this can scarcely be regarded as the correct view. The family is repre- sented in the two halves of the northern hemisphere (in the eastern by sElurus], in both of which it dates from the Pliocene, and, as it is unknown in South America till the Plistocene or late Pliocene, it is evidently one of northern origin ; the American forms having probably attained their maximum development in the Sonoran region. Much the same is the case with regard to the skunks among the Mustelidcz ; these being probably an original Sonoran type which has spread northwards into the Holarctic region and southwards into the Neogaeic realm. Of these, the single species of climbing skunk (Sptlogale) is mainly Sonoran, although it also enters the Transition zone of the Holarctic, and likewise Central America. Of the other members of the group, the typical skunks (Mephitis) range from Hudson Bay to Guate- mala ; while the allied genus Conepatus is found from Texas to Patagonia. In the same family the American badgers (Taxided), although ranging well into the Holarctic, are regarded by Dr Merriam as of Sonoran origin. These badgers, it may be ob- served, differ from the true badgers of the Old World by the form and characters of their cheek-teeth, the last upper molar being proportionately much smaller. Turning to the rodents, the well-known prairie -marmots ( Cynomys], which occupy a position intermediate between the true marmots and susliks, are regarded by Dr Merriam as of Sonoran origin, although they extend into the Holarctic. In the Muridce the peculiar cricetine genus Rhithrodontomys — which, together with the allied South American Rhithrodon, differs from the other members of the sub-family to which it belongs by its grooved upper incisors — appears to be restricted to the region under con- sideration. The white-footed mice (Sitomys), although distributed over the whole of the New World, seem to attain their maximum specific development in the Sonoran, to which the two sub-genera 366 THE SONORAN REGION. [CHAP. Onychomys and Oryzomys are restricted. Yet their near alliance to the Old World hamsters indicates that the group must have had a northern origin, although the genus may have attained its present distinctive features within the Sonoran area. The genus Sigmodon, which differs from the last in the pattern of the molar teeth, and is represented solely by the rice-rat, does not range north of the Sonoran region, although extending into South America as far as Ecuador. The wood-rats (Neotoma), in which the molars simulate the prismatic appearance of those of the voles, .are also largely Sonoran, although they extend into the Canadian FIG. 72. FACE OF Geomys bursarius, SHOWING GROOVED UPPER INCISORS AND OPENINGS OF CHEEK-POUCHES. FIG. 73. FACE OF 7"komomys talpoides, SHOWING SMOOTH UPPER INCISORS AND OPENINGS OF CHEEK-POUCHES. sub-region of the Holarctic, where they are represented by a distinct sub-genus (Teonoma). The round-tailed musk-rat of Florida (Neofiber) is an exclusively Sonoran type, although it is regarded by Dr Merriam merely as a sub-genus of Microtus. Highly characteristic of the region are the pouched rats, consti- tuting the genera GeomyidcR. Of these, the typical genus Geomys * extends northwards into the Transition zone and southwards into Central America ; while the nearly-allied Thomomys> in which the upper incisor teeth are smooth instead of grooved, penetrates into the Canadian sub-region of the Holarctic, although unknown 1 Subdivided into eight genera by Merriam, North American Fauna, Part viii., Washington (1895). X.] ITS FAUNA. 367 in Central America. Both these genera are represented in the Pliocene of the Sonoran area. In the same family the three genera of kangaroo-rats known as Dipodomys, Perodipus, and Microdipodops appear to be confined to the region ; and the same is the case with the allied genus Hderomys, although Perognathus passes northwards into the Transition zone. FlG. 74. UNDER SURFACE OF LEFT FORE-FOOT OF Geomys. In the Ungulata, the deer belonging to the peculiar American genus Cariacus are very abundant in the Sonoran region (where those of the typical genus Cervus are entirely wanting), although they also range into the Canadian sub-region of the Holarctic, and extend right through South America. Since, however, they are wanting in the earlier Tertiary deposits of the latter area, as they are at all epochs in the Old World, there can be little hesitation in regarding them as essentially Sonoran types. Even more decidedly is this the case with the prongbuck (Antilocapra), the sole type of the family Antilocapridce, which is distinguished from the Bovidce. by the horn-sheaths of the males being branched and periodically shed from their bony supports. Although the prongbuck pene- trates a considerable distance into the Canadian sub-region of the Holarctic, its true home is the prairie-district of the Sonoran lying 368 THE SONORAN REGION. [CHAP. to the westward of the Mississippi. Possibly a small deer-like animal from the Tertiaries of the same area known as Cosoryx, may have been the ancestral stock of the prongbuck. Lastly, the peccaries (Dicotyles), which are now chiefly South American, appear to have been originally Sonoran types which have migrated south- wards ; their fossil remains being common in the Tertiaries of the United States, whereas they are unknown in South America before FlG. 75. HEAD OF MALE MULE-DEER (Cariacus macrotis). the Plistocene. Their near affinity to the earlier Tertiary pigs of the Old World indicates that at a more remote date they spread from a more northerly starting-point. With regard to the armadillo (Tatusia) found in the Sonoran, this is clearly a very recent immigrant from the Neogseic realm ; and although opossums (Didelphys) were abundant in North America during the early portion of the Tertiary epoch, it is not improbable that the same explanation will hold good for their existing Sonoran representatives. X.] ITS FAUNA. 369 The following list includes such exclusively New World genera of mammals (apart from bats) which are represented in the Sonoran area; those which may be regarded as more or less nearly confined to this region being printed in italics. To appre- ciate fully the significance of this list, reference must, however, be made to the series of Holarctic genera given on p. 360, which FIG. 76. HEAD OF MALE PRONGBUCK (Antilocapra americana). are more or less completely restricted to the Canadian sub-region of that great region, and the intervening Transition zone. The Sonoran list is as follows, viz. : — Insectivora. SORICID^E. Notiosorex. Also Central America. Blarina. Enters Canadian sub-region of Holarctic. L. 24 370 THE SONORAN REGION. [CHAP. Insectivora (cont.). TALPID^E. Scalops. Enters Transition zone. Scapanus. Enters Canadian sub-region of Holarctic. Carnivora. PROCYONID^E. Bassariscus. Enters Transition and Central America. Procyon. N. to S. America. Nasua. Also South American. MUSTELID^E. Spilogale. Enters Transition and Central America. Conepatus. Texas to Patagonia. Mephitis. Extends into Canadian sub-region and Central America. Taxidea. Enters Holarctic. Rodentia. Cynomys. Extends into Holarctic. MURID^E. Rhithrodontomys. Sitomys. The whole of America. Sigmodon. Southwards to Ecuador. Neotoma. Ranges into Holarctic. Neofiber. GEOMYID^E. Geomys. Extends into Transition zone and Central America. Thomomys. Ranges into Canadian sub-region. Dipodomys. Perodipus. Microdipodops. Perognathus. Ranges into Transition zone. Heteromys. Ungulata. A NTIL OCA PRIDJE. Antilocapra. Ranges into Canadian sub-region. X.] EXTINCT MAMMALS. 37! Ungulata (cont.). Cariacus. Greater part of America. DlCOTYLID^E. Dicotyles. Also South American. Edentata. DASYPODID^E. Tatusia. South American. Marsupialia. DlDELPHYID^E. Didelphys. South American. Although the Transition zone undoubtedly forms an unsatis- factory item in regard to the distinctness of the Sonoran region, yet when we look at the difference of its mammalian fauna as a whole from that of the Canadian sub-region of the Holarctic, and the close similarity between the latter and the fauna of northern Europe and Asia, there can be but little hesitation in regard to the acceptance of Dr Merriam's view that the Sonoran is a valid zoological region of the Arctogaeic realm. In a previous chapter the groups of mammals, both living and extinct, confined to the eastern division of the Arctogaeic realm have been already noticed, while oraupa of in the present one reference has been made to such existing types as are restricted to the western half of Western , 6 /F . ., Arctogaea. of the same realm. It now remains to consider briefly some of the leading extinct groups which are found only in the latter area ; and the consideration of these comes most appro- priately here, seeing that the majority of these peculiarly American types are of Sonoran origin, a large number of their remains having been obtained from the States of New Mexico, Kansas, Nebraska, and Dakota, which lie within that region, or from Colorado, Wyoming, and Montana, which are situated within the Transition zone. Although, in common with the higher Primates, lemuroids are now quite unknown in North America, they were well represented there during the Puerco epoch of the lower Eocene by three families. The first of these — the Chriacida — includes animals 24—2 372 THE SONORAN REGION. [CHAP. X.] EXTINCT MAMMALS. 373 having the same number of teeth as the allied Tertiary European family Adapidce, but all characterised by their more primitive structural features. Indeed these early lemuroids appear to present considerable resemblances to the creodont Carnivora, and differ from all the other members of the sub-order to which they belong by the great elongation of the bony symphysis con- necting the two branches of the lower jaw at the chin. Several other genera, in addition to the typical Chrtacis, are assigned to this family. The second group is that of the Anaptomorphidce, which is represented in the Puerco Eocene by a genus known as Indrodon, and in somewhat higher beds by the typical Anapto- morphus. Although in other respects coming closer to existing types than is the case with the Chriacidcz, the present family is broadly distinguished by the tritubercular structure of the upper molar teeth. The third peculiar North American family of the lemuroids is that of the Mixodectidcz, typically represented by Mixodectes of the Puerco Eocene. Among the extinct creodont Carnivora there are two families apparently restricted to the Tertiaries of North America, namely, the Miacidce, and the Mesonychidcz, the former of which presents such strongly marked affinities to the modern Carnivora that it is frequently assigned to that group. The second family, on the other hand, as represented typically by the genus Mesonyx of the Uinta or lowest Oligocene, is characterised by the very simple structure of the whole series of cheek-teeth, which are not unlike the pre-molars of some of the higher carnivores. One of the species of the typical genus attained dimensions as large as those of a bear. In the widely distributed family Hyanodontida, an exclusively North American genus is Patriofelis, which is regarded as a specialised offshoot from Oxyana. Among the ungulates there are several extinct families con- fined to North America. In the group forming a transition between the pigs and the ruminants there is first of all the family of the oreodonts (Cotylopida), which make their appearance in the middle Oligocene, and continue to the Miocene and lower Pliocene1. These ungulates, which were allied to the genus 1 By American geologists the term Oligocene is not generally used, so that the whole of the Tertiary strata are classed as Plistocene, Pliocene, Miocene, 374 THE SONORAN REGION. [CHAP. 1 X.] EXTINCT MAMMALS. 375 Ancodus, common to the Tertiaries of both hemispheres1, and were represented by a large number of generic types, have crescentic columns to the short-crowned cheek-teeth, the upper molars usually carrying four such columns ; while the lower canine is approximated to the incisors, its usual form and function being assumed by the first pre-molar. The last upper pre-molar is simpler than the molars ; and while the feet have usually four toes each, in the typical genus Cotylops a rudiment of the thumb is retained in the front pair, as in Ancodus. In Cotylops and most of the other genera the molars of the upper jaw have but four columns, but in Protoreodon there are five ; — a feature serving to connect the family with the Anthracotheriida, from which group the oreodonts are probably descended. A nearly-allied but more specialised family is that of the Agriochceridce, as represented by the genus Agriochcerus* , of the upper and middle Oligocene, in which the toes were developed into claws, instead of being incased in hoofs. Here it may be mentioned that while the peccaries (Dicotylida) are now exclusively New World types, and the pigs (Sutdce) restricted to the Old World, the Tertiaries of both and Eocene. Introducing the former term, the series may be approximately classified as follows, viz. : PLISTOCENE. Equus Beds. Eqtius^ Elephas primigenius. PLIOCENE (Upper). Blanco Series. Pliauchenia. (Lower). Loup-Fork. Protohippus, Hipparion. MIOCENE. Deep River. Anchitherium^ First Mastodons. OLIGOCENE (Upper). John Day. Miohippus, Ancodus. (Middle). White River. Agrwchcerus, Titanotherium. (Lower). Uinta. Amynodon, Mesonyx. EOCENE (Up. and Mid.). Bridger. Pachynolophus, Paltzosyops. (Lower). Wahsatch. Hyracotheritim, Coryphodon. (Lowest). Puerco. Neoplagiaulax , Poly mastodon. In this series the Deep River beds are identified with the European Miocene (supra^ p. 117) by the presence of Anchitheritim, and the first appear- ance of Mastodon ; while the existence of Ancodus in the John Day and Upper White River beds correlates them with the Upper and Middle Oligocene, Hyracotherium and Coryphodon serving to identify the Wahsatch beds with the Lower Eocene. 1 Supra, p. 161. 2 Equal Artionyx. 376 THE SONORAN REGION. [CHAP. hemispheres contain intermediate types such as Hyotherium and Chcerohyus which are apparently the ancestral stock of both families. FIG. 79. FRONT VIEW OF RIGHT HIND FOOT OF Agriochcerus. Another very peculiar type of North American Tertiary ungu- lates is represented by Protoceras, from the upper division of the White River Oligocene, which forms a family (Protoceratidce} by itself. In these creatures the feet approximate to the ruminant type; but the skull, as shown in the accompanying illustration, FIG. 80. SKULL OF Protoceras, WITHOUT THE LOWER JAW. has at least two pairs of large bony processes, probably covered in life with horns, and a pair of large upper tusks, in both of which respects it exhibits a curious parallelism with the perissodactyle X.] EXTINCT MAMMALS. 377 ungulates. No trace of these singular artiodactyles has hitherto been detected in the Old World. The Camelidce. seem to have been primarily a North American family, which originated in the Sonoran region, and of which one branch (Lama) subsequently migrated south, while the other (Camelus) crossed Bering Strait into the Old World. In the upper Pliocene there occurs Pliauchenia, with only three lower pre-molars, and in the lower Pliocene Loup-Fork beds Procamelus with four of these teeth ; while the earliest representative of the family is Leptotragulus of the Uinta Oligocene. In the perissodactyle section of the same order the family Titanotheriidce is mainly North American, although, as stated on page 107, a representative of the typical genus Titanotherium has been discovered in the Tertiaries of the Balkans. Titanotherium includes huge rhinoceros-like animals, with low-crowned molar FIG. 8r. RIGHT UPPER MOLAR TOOTH OF Palaosyops, teeth of the type of those of Chalicotherium, and frequently having the nasal region of the skull surmounted by large bony protuber- ances. The genus is characteristic of the Uinta and the lower division of the White River Oligocene. An earlier type of the same family is typified by the smaller and less specialised hornless animals from the Bridger Eocene, known as Palaosyops, which, together with certain allied forms, constitute a peculiar sub-family confined to America. This family, like the camels, appears there- fore to have originated in the Sonoran region, whence a few representatives wandered eastwards into the Old World. In the generalised ungulate sub-order termed Amblypoda, of which the lower Eocene coryphodons were the earliest representatives, North America possesses an absolutely peculiar THE SONORAN REGION. [CHAP. family in the UintatheriidcR. These were huge, somewhat ele- phantine ungulates, with five short toes to each foot, long tusk- like canine-teeth in the upper jaw, and the skull surmounted with three pairs of large bony protuberances ; their molar teeth being a specialised form of the Coryphodon type. They occur in the middle division of the Bridger, or the one above the zone yielding Coryphodon, and may accordingly have been the descendants of that genus. These uintatheres appear to have been restricted to the " Bad Lands " of the Sonoran region and adjacent districts of the Transition zone. FlG. 82. EXTREMITY OF SKULL OF UintCltherium , TO SHOW UPPER TUSKS. Passing by certain other forms of less interest, attention may be directed to a peculiar group of aberrant mammals forming the Tillodontia, which appear to be mainly North American, and of which the serial position cannot be precisely determined. They are restricted to the Eocene, and seem to combine the characters of the modern Ungulata, Rodentia, and Carnivora. In the genus Anchippodus, forming the type of the family AnchippodontidcB, the skull approximates to that of a bear, the cheek-teeth are of an ungulate type, and there is a pair of large chisel-like incisors (preceded by a small functionless upper pair) in each jaw, very similar to those of the rodents and hyraces. A second family, X.I LIST OF THE FAUNA. 379 Psittacotheriida, is represented by the genera Psittacotherium and Calamodon, in which the cheek-teeth grew permanently, instead of developing roots. The following list exhibits the chief families or minor groups of characteristically North American mammals, which are either entirely wanting or but sparingly represented in the Old World ; those that are extinct being indicated by a t, and the absolutely characteristic forms being printed in italics. Primates. t Anaptomorphidce. Anaptomorphus. t Mixodectida. Mixodectes. t Chriacidce. Chriads. Carnivora. Procyonidae. Represented in the Old World only by ^Elurus. t Miaridce. Miacis, Didynrictis. t Mesonychidcz. Mesonyx. fHyaenodontidse. Patriofelis. Rodentia. Haplodon tidce. Geomyidce. Ungulata. Dicotylidce. Represented by ancestral types in Tertiaries of E. Hemisphere. t Cotylopida. Cotylops, Mesoreodon, Protoreodon. t Agriochceridce. Agrioch&rus. t Protoceratida. Protoceras. Camelidae. t Pliauchenia, \Procamelus, M^eptotraguius. Antilocaprida. Antilocapra, t Cosoryx. t Titanotheriidae. Palceosyops, Limnohyops, Telmatotherium. t Uintatheriidce. Uintatherium. Tillodontia. t Psittacotheriida. Psittacotherium, Calamodon. t A nchippodontidce. A nchippodus. 380 THE SONORAN REGION. [CHAP. In an earlier chapter1 a list has been given of the leading Distinctness mammalian families common to the two divisions of the Region. of Arctogaea, and since in the foregoing chapter it has been shown that many of the peculiar American families are more or less intimately related to some of those common to the two areas, it is manifest that throughout the Tertiary period eastern and western Arctogaea must have had a land-connection towards the north, so that there was an interchange of the fauna of the more northern districts. Those American types which penetrated as far south as what is now the Sonoran area would, however, naturally tend to become isolated, and thus develop into the families which may be regarded as characteristic of that region. So far, therefore, from this area being merely a part of a so-called Nearctic region, there are indications that it was differ- entiated from the Holarctic at a time when the existing zoological regions of the eastern half of the Arctogaeic realm were still unde- fined. Indeed from the community of the Pliocene fauna of southern Europe, Asia Minor, Persia, northern India, and south China, it seems probable that the only divisions of the Arctogaeic realm that could have been attempted would have been into (i) a Sonoran region, (2) a Holarctic region, comprising the northern districts of America, Asia, and Europe, (3) what may be termed a Mediterraneo-Oriental region, including southern Europe, north Africa, and the whole of southern Asia ; and (4) a Malagasy region, which would then, or perhaps somewhat earlier, have included Ethiopian Africa. As to the amount of interchange which took place during Dual Origin Tertiary times between the mammals of the eastern of Groups. and western divisions of Arctogaea, and as to whether similar generic types may have been developed independently in the two areas, it is almost impossible to arrive at any satisfactory conclusion. The suggestion that Equus has thus been independ- ently evolved in the two areas, has been already mentioned2, and this idea receives support from some very remarkable observations recently made on the invertebrates inhabiting certain European and North American caves. 1 Sttpra, p. 176. 2 Supra, p. 168. X.] DUAL ORIGIN OF GROUPS. 381 With a quotation from Mr G. H. Carpenter's interesting paper1 on this subject, the present volume may be fitly closed. After describing the inhabitants of the Mitchelstown Cave, in Ireland, the author writes that the spring-tail (Lipurd] "is hardly to be separated from a species found in the caves of Carniola, and the Sinella is almost identical with one inhabiting the caves of North' America ; while the spider is apparently the same as a cave- dweller from the Mediterranean district of southern France, which probably occurs in the North American caverns also. Had we to do with animals of the upper fauna, these results, though highly interesting, would not be without parallel in species already known.... But the occurrence of cave-dwelling species with so wide a range is a truly remarkable phenomenon. The caves cannot be of any great geological age. Any possible geographical connection which would permit the migration of subterranean animals between southern Europe and Ireland, or between Ireland and North America, seems altogether out of the question within any period during which the fauna can have been specifically identical with that of the present day. The only conclusion is that from ancestors, presumably of the same genus, which took to an underground life in such widely-separated localities, the similar conditions of the caves have evolved descendants so similar that, when compared, they cannot or can hardly be specifically distin- guished from each other. Should the identifications stand the test of a comparison of types, we shall have proof that the inde- pendent development of the same species, under similar conditions, but in widely-distant localities, has taken place. It must be granted, however, that cave-conditions are so marked and excep- tional, that it might not be safe to argue from them as to what may have occurred in the upper world." Although the author of this passage is perfectly correct in his statement that there is a vast difference between cave-life and open-air life, yet if animals which appear to belong to one and the same species can be proved to have had a dual origin in the one case, it can scarcely be considered impossible that similar instances may occur in the other. And if such dual origins exist among 1 Irish Naturalist, Vol. IV. pp. 25 — 35 (1895). 382 THE SONORAN REGION. [CHAP. X. species, there is surely no reason why they should not occasionally occur in the case of genera. It would, therefore, seem by no means improbable that the species of the genus Equus which inhabited the eastern and western halves of the northern hemi- sphere during the close of the Tertiary period may have been evolved from the closely-allied but separate ancestral equine stocks. The matter does not, however, by any means end here. In an earlier chapter1 it has been shown that the same species of a genus of fish (Galaxias) occurs in countries so remote from one another as New Zealand and Australia on the one hand, and Patagonia on the other. With the evidence of the cave-animals before us, is it absolutely impossible that these apparently iden- tical fishes can have been evolved independently of each other ? Should this be so, it will engender increased caution in drawing any inferences as to former land-connection from the evidence of single animals. But such instances of independent evolution, if they do occur, must be of extreme rarity, and will in no case interfere with deductions drawn from the presence of a number of allied species or genera of animals in distant countries. 1 Suprh, p. 125. Recently another species has been described from South Africa. See Steindachner, SB. Ac. Wien, vol. ciii. p. 460 (1894). APPENDIX. LIST OF WORKS AND PAPERS MOST FREQUENTLY REFERRED TO IN THIS VOLUME. 1. ALLEN, J. A. The Geographical Distribution of Mammals. Bull. U. S. Geol. Survey, vol. IV. nos. 2 and 4, pp. 313—376, (1878). 2. The Geographical Distribution of North American Mammals. Bull. Amer. Mus., vol. IV. pp. 199—243, (1892). 3. ANONYMOUS. Antarctica; a Supposed Former Southern Conti- nent. Natural Science, vol. III., pp. 54—57, (1893). 4. The Nearctic Region and its Mammals. T. c. pp. 288 — 292. 5. BEDDARD, F. E. A Text-Book of Zoogeography. (Cambridge Natural Science Manuals}. Cambridge, 1895. 6. BLANFORD, W. T. The African Element in the Fauna of India. Ann. Mag. Nat. Hist. ser. 4, vol. XVIII., pp. 277 — 294, (1876). 7. Note on the "Africa- 1 ndien" of A. von Pelzeln, and on the Mammalian Fauna of Tibet. Proc. Zool. Soc. 1876, pp. 631—634. 8. Anniversary Address to the Geological Society. Proc. Geol. Soc. 1890, pp. 43—110. 9. The Age of the Himalayas. Geol. Mag. decade 3, vol. IX., pp. 161 — 168, (1892). 10. — On a Stag (Cervus thoroldi} from Tibet, and on the Mammals of the Tibetan Plateau. Proc. Zool. Soc. 1893, pp. 444-449. 11. BOURNS, F. S., and DEAN, C. W. Preliminary Notes on Birds and Mammals collected in the Philippine Islands. Occasional Papers Minnesota Acad. vol. I., pp. I —64, (1894). 12. BULMAN, G. W. The Effect of the Glacial Period on the Fauna and Flora of the British Islands. Natural Science, vol. ill., pp. 261 — 266, (1893). 384 APPENDIX. 13. CARPENTER, G. H. Nearctic or Sonoran ? Natural Science, vol. v., pp. 53—57, (1894). 14. EVERETT, A. H. A Nominal List of the Mammals inhabiting the Bornean Group of Islands. Proc. Zool. Soc. 1893, pp. 492—496. 15. FORBES, H. O. The Chatham Islands; their Relation to a Former Southern Continent. Supplemental Papers R. Geogr. Soc, 1893, pp. 607—637. 1 6. HEDLEY, C. On the Relation of the Fauna and Flora of Australia to those of New Zealand. Natural Science, vol. III., pp. 187— 191, (1893). 17. HEILPRIN, A. The Geographical and Geological Distribution of Animals. International Scientific Series, London, 1887. 1 8. HUXLEY, T. H. On the Classification and Distribution of Alectoromorphas and Heteromorphas. Proc. ZooL Soc. 1868, pp. 294—319. 19. MERRIAM, C. H. The Geographic Distribution of Life in North America, with special reference to the Mammalia. Proc. Biol. Soc. Washington, vol. VI I., pp. i — 64, (1892). 20. Laws of Temperature Control of the Geographic Dis- tribution of Terrestrial Animals and Plants. Nat. Geogr. Mag. vol. VI., pp. 229 — 238, (1894). 21. NEHRING, A. Ueber Saugethiere von den Philippinen, nament- lich von der Palawan-Gruppe. Sitzber. Ges. Naturf. Berlin, 1894, pp. I79—I93- 22. OGILBY, J. D. Catalogue of Australian Mammals. Sydney, 1894. 23. OSBORN, H. F. The Rise of the Mammalia in North America. Stud. Biol. Lab or at. Columbia College, Zoology, vol. I., art. 2, (1893). 24. , and EARLE, C. Fossil Mammals of the Puerco Beds. Bull. Amer. Mus. vol. VIL, pp. i — 70, (1895). 25. SCHARFF, R. F. On the Origin of the Irish Land and Fresh- water Fauna. Proc. Irish Acad. ser. 3, vol. III., pp. 479 — 485, (1894). 26. SCLATER, P. L. On the General Geographic Distribution of the Members of the Class Aves. Journ. Linn. Soc. Zool. vol. II., pp. 130—145, (1858). 27. The Geographical Distribution of Mammals. Man- chester Science Lectures, ser. 5 and 6, pp. 202 — 219, (1874). 28. W. L. The Geography of Mammals. Geographical Journal, 1894, 1895. APPENDIX. 385 29. SHARPE, R. B. On the Zoo-Geographical Areas of the World, illustrating the Distribution of Birds. Natural Science,\o\. III., pp. 100—108, (1893). 30. THOMAS, O. T. The Mammals of the Solomon Islands. Proc. Zool. Soc. 1888, pp. 470—484. 31. Preliminary Diagnosis of New Mammals from Northern Luzon, collected by Mr J. Whitehead. Ann, Mag. Nat. Hist. ser. 6, vol. XV. pp. 160 — 164, (1895). 32. WALLACE, A. R. The Geographical Distribution of Animals. London, 1876, 2 vols. 8vo. 33. Island Life. London, 1880, 8vo. 34. What are Zoological Regions? Nature, vol. XLIX., pp. 6 10 — 613. 35. The Palaearctic and Nearctic Regions compared as re- gards the Families and Genera of their Mammalia and Birds. Natural Science, vol. IV., pp. 433 — 445, (1894). 36. ZITTEL, K. A. VON. Die geologische Entwickelung, Herkunft, und Verbreitung der Saugethiere. Sitzber. bayer. Akad. vol. xxill., pp. 137—198, (1893). L. 25 INDEX. Aard-varks, 187, 249, 256, 268 Aard-wolf, 235 Abderites, in Abderitidcz, 1 1 1 Acaremys, 90 Acdestis, no Acomys, 198 Ac other uhim.) 192 Acrobates, 37 Adapis, 181, 191, 219 Addax, 245, 324, 358 Adelphomys, 91 ^ 157, 191, 202 $) 321 212, 275 sEpyceros, 245 sEpyornis, 222 sEpyprymnus, 36 Age of animal groups, 7 Agriochceridiz, 375 Agriochceriis, 375 Agutis, 89, 137 Alactaga, 323, 351 ^&w, 164, 315 Amblotherium, 53 Amblypoda, 173, 377 Amblyrhiza, 87, 137 Ameghino, on age of Santa Cruz fauna, 68 Ammodorcas, 245 AmphictiSy 191 Amphicyon, 158, 191 .4 ?«/>// icy nod on ,193 Amphidozotherium, 156, 191 Amphilestes, 53 Amphiproviverra, 109 Amphisbcend) 132 Amphisbcenidtz, distribution of, 132; fossil, 132 AmphitheriidcB, 52 Amphitherium, 53 itragulus, 1 94 ^? nalcitheriu m, 105 Anaptomorphidce, 373 Anaptomorphus, 373 Anchilophus, 166, 192, 193 Anchippodontid&i 378 AnchippodttS) 378 Anchithei'ium, 166, 196 Ancodus, 161, 193, 203, 375 Ancyclotherium, 172, 200 Anoa, 47, 164 Anomaluridce, 237 Anomalurus, 238 Anomodontia, 151 AnoplotheriidcZ) 185 Anoplotherium, 185, 193 Anops, 132 Antarctica, 56, 128; extent of, 129 Antarctogsea, 26 Anteaters, 101, 136 ^4 ntech inom ys, 3 9 Anthracotheriida, 161, 192 Anthracotherium,) 163, 192, 193, 194,. 203 Anthropopithecus, 180, 202, 231, 256 Antillia, 139 Antillean sub-region, 136, 137 Antilocapra, 367 Antilocapridce, 367 Antelopes, African, 242 Amirosorex, 272 Africa connected with S. America, 127 Africa, route of migration into, 256 Aplin, Mr, on mammals crossing rivers, 14 Aquilonian region, 26 Arabia, South-eastern, fauna of, 262 Arapaima, 134 Arch&lurus, 157 Archceoniys, 91 INDEX. 387 Arctictis, 273 Arctic sub-region, 360 Arctic zone, 310 Arctocyon, 159 Arctogale, 273 Arctogsea, 25, 26, 27, 144, 210 Arctogsea, Eastern, 179 Arctogasan region, 26 Arctogaeic fauna, features of, 147 Arctogseic realm, 27, 144 Arctotherium, 72 Arctomys, 159, 267, 351 Arc t onyx, 276 Area, 2 Armadillo, giant, 136 Armadillos, 94, 368 Artionyx, 375 Arui, 35« Arvicanthis, 239 Asses, 359 Astrapotheria, 81 Astrapotheriutn, 81 Atherura, 278 Atlantic, its recent origin, 24 Atlantosaurus, 35, 151 Auk, 347 Anlacodus, 91, 240 Australia, how it received its fauna, 54; its connection with S. America, 55» 125; its relations with New Zealand and other islands, 58, 59 Australian region, 25, 27, 31 Austro-Columbian region, 25 Austro-Malayan region, 27, 28, 45 Avahi, 217 Awantibo, 256 Aye-aye, 219 Babirusa, 47 Bachitherium, 164, 192 Badgers, 321 Balanoptera, 68 Banded anteater, 40 Bandicoots, 38 Banteng, 278 Barbets, 254 Barriers to dispersal, 10, 14 Bassariscus, 73, 136, 364 Bathyergus, 23 r, 239 B atomy s, 278 Bats, Notogseic, 42 ; Neogseic, 70 Baur, Dr, on relations of Galapagos islands, 141 Bdeogale, 235 Bears, American, 318; their absence from Ethiopian Africa, 258 Beavers, 313, 318 Beddard, on earth-worms, 126, 311; on Fernando Noronha, 140 Bering Strait indicates a line of con- nection between Asia and America, 178? 337 Bettongia, 30 Birds, Malagasy, 254; Ethiopian, 254 Bison, 314 Black-buck, 278 Blanford, Dr, on permanency of conti- nents, 24 ; on zoological regions, 27 ; on a former connection between the southern continents, 128; on con- nection between Africa and S. America, 128; on survival of old forms in Madagascar, 131; on the birds of Madagascar and Africa, 254 ; on Oriental sub-regions, 266 ; on the fauna of India, 267 ; on land connection between Ceylon and Malaysia, 293 ; on the Tibetan fauna, 354 Blarina, 364 Blastomeryx, 74 Blesbok, 244 Boa, 132 Boas, distribution of, 132 Bontebok, 244 Boreal region, 146, 309, 360; zone, 360 Borhycena, 109 Borneo, its fauna, 294 Bos, 165, 206, 278, 314; depressicornis, 47, 164; mindorensis, 47 Boselaphus, 206 Both riospondyhis, 152 Bothremys, 131 Bothriceps, 152 Bourbon, 213 Bovidce, 164, 186, 196; Holarctic, 314; Oriental, 279 Brachydiastematotherium, \ 70 Brachyurns, \ 36 Brachyteles, 136 Bradypodida, 100 Brady pus, 100 Bramatherium , 186, 204 Brazilian sub-region, 135 British Isles, their connection with the Continent, 339 Brontornis, 60 Bubalis, 206, 244, 358 Budorcas, 324 Buffalo, 279 Bulman, Mr, on the submergence of Britain, 350 25—2 388 INDEX. Buphaga, 254 Burmese sub-region, 267 Buried river-channels of Britain, 339 Cadurcotherium^ 82, 170, 192 Ccenolestes, no C¬heriidce, 163, 192 Ccenotherium, 163, 192, 193 Calamodon, 379 Callithrix, 136 Caloprymnus, 36 CamelidcE, 163, 185, 326, 377; Neo- gseic, 74 Camelus, 163, 185, 204, 377 Camels, 326 Campos of Brazil, 65 Canadian sub-region, 360 Cane-rat, 91 Canidce, Oriental, 274; Neogaeic, 72 Cam's, 195, 202, 235, 274; antarc- ticus, 140; dingo, ^i;frocyonides, 355 Cannabateomys, 91 Capra, 206, 230, 280, 324, 359, 360 Capreolus, 352 Capivara, 88 Capromyida, 90 Capromys, 91, 137 Caracal, 249 Cardiotherium, 89 Carcttochelyidcz, 62 Carettochelys, 62 Cariacus, 74, 344, 367 Carioderma, 96 Carnivora, Arctogseic, 157; Holarctic, 312, 321; Neogaeic, 71; North American, 341 ; Sonoran, 364 Carpenter, Mr, on cave-faunas, 381 Carpincho, 88, 136 Carpotnys, 278 Carterodon, 91 Castor, 313 Castorida, 159, 313 Castoroides, 87, 138 Castoroidida, 87, 138 Cassowaries, 48, 58 Casuarius, 48 Catamarca beds, 67 Catonyx, 105 Caves of Lagoa Santa, 66 Ctf^a, 88 Cavies, 88 Caviidcz, 88 Cave-faunas, 381 Caxomistle, 136 Cebidce, 69 Cebochczrus, 192 Celebes, its fauna, 47; an anomalous island, 49 Cewas, 280 Centetes, 220 Centetidce, 131, 220 Central American sub-region, 136 Central Asian sub-region, 351 Centres of evolution, 2 r Cephalogale, 191 Cephalophus, 244 Ceratodns, 133 Ceratophora, 120 Cercocebus, 231 Cercoleptes, 72 Cercomys, 91 Cercopithecida, 231, 269; distribution of, 180 Cercopithecus, 231 Cernaysian Fauna, 153 Cervicapra, 245 Cervicaprince, 244, 245 Cervidce, 164; absence of in Ethiopian region, 230, 258; Holarctic, 315 ; Oriental, 280 Cervulus, 198, 281 CVrzw, 164, 280, 326, 335'.358> 359> 367; eustephanus, 352; titnoriensis , 46; xanthopygus, 351 Cestracion, 63 Cetaceans of American Tertiary, 68 Cetotherium, 68 Ceylon, when separated from India, 293 Ceylonese sub-region, 266 Chcerokyus, 376 Chceromeryx, 203 Chtzropotamida:, 161, 192 Cheer opotamus, 161, 192 Cheer optis, 39 Characiniidce, distribution of, 134 Chcetomys, go, 136 Chalicomys, 159, 194, 196, 313 Chalicotheriidce, 196, 200 Chalicotherium, 170, 192, 200 Chameleons, distribution of, 222 Chamois, 324 Chelonians of Notogsea, 62 Chevrotains, 164, 281 Chimarrogale, 272 Chilian sub-region, 135 Chimpanzees, 180, 231, 256 Chinchillas, 89, 136 Chinchillidee, 89 Chipmunks, 313 Chiromyidce, 181 Chiromys, 219 Chiropodomys, 278 INDEX. 389 Chiroiiectes, 107 Chiru, 325 Chirnromys, 41 Chlamydophorus, 94, 96, 137 CJilamydotherium, 96 CholcepuS) 10 1 Chriacidie, 371 C/iriacis, 373 Christmas Island, 303 Chromididiz, 134 Chrotomys, 277 Chrysochloridiz, 230, 234 Ckrysochloris, 230, 234 Civets, 182; Oriental, 273 Clccnodon, 159 Claviglis, 238 Coatis, 72, 365 Cobus, 200, 206, -245 Cwlogenys, 89, 136 Colics, 254 Collide?, 254 Colobus, 231 Colymbus, 347 Comoro Islands, 213 Conepatus, 73, 365 Condylura, 341 Conihtrus, 41 Connection between Africa and S. America, 127 ; Australia and S. America, 125 Connochcetes, 244 Continental islands, 10 Continents, permanency of, 22 Cooke, Mr, on S. American molluscs, 124; on Antillean molluscs, 139; on N. American land Mollusca, 3n Cope, Prof, on N. American fossil cetaceans, 68 Coral his, 132 Coryphodontidce, 174 Coryphodon, 174, 378 C0.SWJ.*-, 368 Cotton-rat, 88 Cotylopidce, 373 Cotylops, 375 Coypu, 91, 137 Crateromys, 277 Cricetince, 160, 183, 239 Cricetomysy 239 Cricetus, 160, 192, 196, 322 Crocidura, 195, 272 Crocodilus porosus, 62 Crossarchus, 235 Crossopus, 319 Cryptobranchus, 35 Cryptodira, distribution of, 8 Cryptoprocta, 220 Ctenodactylidcz, 90 Ctenodactylus, 91, 212, 240, 323, 357, 358 Ctenomys, 90 Cuniculus, 314 Cuscuses, 36, 46 Cusimanse, 235 Cycloturus, 102 Cyncelurus, 202, 234, 249, 272 Cynic tis, 235 Cynodictis, 158, 191 Cynodon, 191, 193 Cynopithecus niger, 47 C>», 235, 274 Cynogale, 273 Cynomys, 342 Dacrytherium, 185, 192 Dactylomys, 91 Dactylopsila, 37 Dcedicurus, 98 Damaliscus, 244 Damuda-Talchir flora, 153 Dapedoglossus, 134 Daphcenus, 158 Dasornis, 60 Dasymys, 238 Dasyprocta, 98, 137 Dasyproctida, 89 Dasypodidce, 94 Dasypus, 94 Dasyuridce, 39 ; Neogaeic, 108 Dawkins, Prof. B., on the range of the mammoth, 334 Decastis, no Deer, 164, 267, 367 ; Oriental, 280 ; their absence from Ethiopian Africa, 230, 258 Deep-sea deposits in islands, 23 Degu, 90 Dendrogsea, 26 Dendrogale, 270 Dendrolagus, 36 Dendromys, 238 Deomys, 239 Desman, 321 Deserts as barriers, 15 Diadiaphoriis , 80 Diatryma, 60 Dichobumts, 163, 192 Dichodon, 192 Dichodontidce, 185, 192 Dicroceros, 196 Dicotyles, 73, 368 Dicotylida, 74, 375 Dicynodonts, 151 390 INDEX. Didelphyidce, Neogseic, 107 Didelphys, 51, roy, 192, 193, 194, 368 Dingo, 42 Dinocyon, 158, 195 Dinomyidce, 89 DinornithidfB, 58 Dinosauria, 151 ; wide distribution of, 8 DinotheriidcB) 187 Dinotherium, 173, 187, 196, 200,206 Dist Holarctic, 312; So- noran, 364 Soriculus, 272 South American region, 26 South America, separated from N. America, 117; invaded by northern animals in the Pliocene, 119 Southern Pateozoic Continent, 153 South African sub-region, 261 Spalacidce, 183, 239, 322 Spalacopus, 91 Spalacotheriidce, $2 Spalax, 183, 322, 358 Spencer, Mr, on buried West Indian river-channels, 138 Spermophilus, 159, 351 Spheniscida, a southern group, 131 Sphenodon, 8, 63 Sphodromys, 90 Spiny anteater, 33 Spiny rats, 136 Spilogale, 365 Sparassodonta, 108 Squirrels, Flying, 237 Star-nosed mole, 341 Station, 2 SteatomyS) 239 Stegodont elephants, 206 Stenodelphis, 1 1 2 Stenoplesictis , 182, 191 Stereornithes, 60 Stereo stern um , 133 Sternothoerus* 131 Sthenurus, 36 StiromyS) 90 Stonestield slate, its marsupials, 52 Strepsiceros, 199, 206, 246 Strut hio, 255; distribution of, 129 Submerged forests of Britain, 339 Sucker-footed bats, distribution of. 13 r Suida:, 184, 375; Oriental, 281 Sumatra, 300 Siiricata, 235 Sus, 184, 203, 242, 267,281; absence of, in Ethiopian region, 230 ; cele- bensis, 48 ; papuensis, 42 Swamp-deer, 280 Swayne, Capt., on the fauna of So- maliland, 261 Syria, formerly connected with Africa, 259 Systemodott) 165 Takin, 324 Talpa, 156, 191, 271, 321 Talpidce, 156, 271, 312; absence in Ethiopia, 230; Sonoran, 364 Tamandua, 102 Tamarao, 47, 279 Tamias, 159, 313 Tatniascvurus ) 344 7\ipiridce, 165; Oriental, 282 Tapirs, 74 Tapirus, 73, 165, 194, 282 Tarsiidce, 181, 270 Tarsipes, 37 TarsiuSi 2 70 ; fuscomanus, 47 Tasmania, 31 Tasmanian devil, 39 Tatusia, 94, 368 Tayra, 73 Tchitrea, 254 Telephomis, 254 Temnocyon, 158 Temperature, influence of, 3 Tenrec, 220 Teonoma, 344, 366 INDEX. 399 Tertiary mammalian fauna of Eastern Arctogoea, 190 Tesiudo, distribution of, 129 Tetraceros, 206, 244, 280 Tetraconodon, 163, 204 Tetracus, 193 Thames Valley, brick-earths of, 335 Theridomyidic, 89, 91, 240 Theridomys, 91, 192 TheropithectiSi 231 Thomas, Mr O., on the fauna of S. E. Arabia, 262 ; on the fauna of Sipora, 303 Thomomys, 366 ThylacimiS) 39 Thylacoleo, 37 Thyropoda, 132 Tibetan sub-region, 352 Tierra del Fuego, 140 Tiger, 272 Tillodontia, 378 Timor, 46 Titanomys, 194 Titanosaurus, 152 Titahotheriidce, 170, 377 Titanotheritim, in the Balkans, 57, I7«>» 377 Tobago, continental, 137 Tolypeutes, 94 Tortoises, Galapagos, 141 Tosca, 66 Toxodon, 82 Toxodontia, 82 Toxodontidce, 83 Toxodontothcrium, 83 Tragelaphus , 199, 246 Tragocerosy 199 Tragulida, 164, 196, 204, 242; Ori- ental, 281 TraguluS) 164, 186, 204, 281 Transition zone, 361 Tratratratra, 219 Tree-porcupines, 136 Triaulacodus , 91, 240 Trichechida, 313 Triconodon, 51 Triconodontida:, 51 Trichosums, 36 Trichys, 278 Triclis, 36 Trigonia, 63 Trinidad, continental, 137 Trilophomys, 238 Tntylodon, 149 Trogonidcz, distribution of, 129 Trogontherium , 333 Tropical region, 140 Trygenycteris, 22 r, 232 Tucotuco, 90 Tup aia, 270 Ttipaiidce, 268, 270 Tuatera, 8, 63 Ture, 348 Tylas, 254 Ty pother iidce, 84 Typolherium, 84 Tyrrhenian sub-region, 357 Uintatheriidcz, 174, 378 Ungulates, E. Holarctic, 324; Neo- gaeic, 73; N. American, 343 ; num- ber of African, 241; Oriental, 278; Sonoran, 367 Uria, 347 Uromys, 41 Uronycteris, 42 Uropsihts, 321 Urotrichtis, 156, 312, 356 Ursidce, their absence from Ethiopian Africa, 258; Holarctic, 321 Ursus, 202, 274, 318; absence of, in Ethiopian region, 236 ; ornatus, Vampire bats, 70 Vandeleuria, 278 Varanus priscus, 62 Vicunas, 74, 136 Viscachas, 136 Vishnutherium, 186, 204 Viverra, 182, 191, 195, 234, 273; tangalunga, 47 Viverricula, 273 Viverridcz, African, 234; an Old World group, 182; Oriental, 273 Voles, 160, 267, 313, 318 W. Africa and Malaysia, affinity of faunas, 292 Wallabies, 36 Wallace, Dr, on extinction of Plistocene fauna, 18; on zoological regions, 25; on the Moluccas, 48 ; on Celebes, 48 ; on origin of Australian fauna, 54 ; on isolation of Australia, 58 ; on S. American fauna, 123 ; on connection between S. America and Australia, 125; on W. Indies, 138; on con- nection of S. and N. America, 153 ; on the past history of Ethiopia, 255; on land connection between India and Africa, 257; on Oriental sub-regions, 266; on relationship between W. African and Malayan 400 INDEX. faunas, 292 ; on Javan fauna, 302 ; on the former union of Malta and Sicily with Italy, 337 Wallace's line, 10 Walruses, 313 Wapiti, 315 Wart-hogs, 256, 268 Water-buck, 245 Water-chevrotain, 242 Water-deer, 356 Water-shrew, 319 West Indian sub-region, 137 West Indies connecting N. and S. America, 139 Western Arctogsea, its extinct mam- mals, 371 \Vhite-footed mice, 88, 342 Wildebeests, 244 Wolverene, 312 Wombats, 38 Wood-hoopoes, 254 Wood-rats, 342, 344, 366 Xenurus, 94 Xeromys, 40, 277 Xerus, 195, 197 Xiphodon, 185, 192, 193 Xotodon, 83 Yak, 314 Zaglossits, 33 Zapus, 342 Zebras, 247 Zoological realms and regions, 25, 27 Zoster ops, 254 CAMBRIDGE: PRINTED BY j. AND c. F. CLAY, AT THE UNIVERSITY PRESS. LOAN PERIOD 1 Sciences Blda.