THE

CAMBRIDGE NATURAL HISTORY

EDITED BY

S. F. Harmer and A. E. Shipley

VOLUME X

MAMMALIA

by F. BEDDARD

Reprint Edition

1958

© 1902, by Macmillan & Co., Limited

Authorized reprint by
Wheldon & Wesley, Ltd. and H. R. Engelmann (J. Cramer)

Printed in Germany

PREFACE

Inasmuch as Sir W. H. Flower and Mr. Lydekker could not profess to treat the Mammalia exhaustively within the limits of nearly 800 pages, in their Introduction to the Study of Mammals, it is obvious that the present volume, which appears ten years later and is of rather less size, can contain but a selection of the enormous mass of facts at the disposal of the student of this group. Thus the chief question for myself was what to select and what to leave aside. It will be observed that I have reduced the pages of this book to conformity with those of other volumes of the series by treating some groups more briefly than others. It has appeared to me to be desirable to treat fully such groups as the Edentata and the Marsupialia, and permissible to be more brief in dealing with such huge Orders as those of the Rodentia and Chiroptera. Lengthy disquisitions upon such familiar and comparatively uninteresting animals as the Lion and Leopard have been curtailed, and the space thus saved has been devoted to shorter and more numerous accounts of other creatures. As there are nearly six hundred genera of living Mammals known to science, omission as well as compression became an absolute necessity. I have given, I hope, adequate treatment from the standpoint of a necessarily limited treatise to the majority of the more important genera of Mammals both living and extinct; but the length of this part of the book had to be increased by the discoveries, which give me at once an advantage and a disadvantage as compared with the two authors whose names I have quoted, of a considerable number of important new types in the last ten years.

Such forms as Notoryctes, Romerolagus, Caenolestes, "Neomylodon," and Ocapia could not possibly have been omitted.

In preparing my accounts of both living and extinct forms I have nearly invariably consulted the original authorities, and have often supplemented or verified these accounts by my own dissections at the Zoological Society's Gardens. My rule has not, however, been invariable in this matter, inasmuch as there exist two recent and trustworthy text-books of Mammalian Palaeontology—Professor Zittel's Handbuch der Palaeontologie, and Dr. A. Smith Woodward's manual, Outlines of Vertebrate Palaeontology, in the Cambridge Biological Series. Where the name of a genus only or its range, or merely one or two facts about it, are mentioned, I have not thought it necessary to go further than these two works. But a good deal has been done even since the appearance of these two volumes which it will be found that I have not ignored.

I have to thank my editors for the trouble which they have taken in the revision of the proofs and for many suggestions. To Professor Osborn, of Columbia University, New York, I am indebted for some kind suggestions. My daughter Iris has assisted me in various ways. Finally, I desire to express my indebtedness to Mr. Dixon and to Mr. M. P. Parker for the care which they have taken in the preparation of the figures which were drawn by them especially for this work.

Frank E. Beddard.

London, February 28, 1902.

CONTENTS

Scheme of the Classification Adopted In This Book

Sub-Class Prototheria (p. [105]).
Order.		Sub-order.		Family.		Sub-family.
MONOTREMATA (p. [106])				Echidnidae (p. [110]). Ornithorhynchidae (p. [112]).
?ALLOTHERIA (p. [96]).
Sub-class Eutheria (p. [116])
MARSUPIALIA (p. [122])		Diprotodontia (p. [128])		Macropodidae (p. [129])		Macropodinae (p. [132]). Potoroinae (p. [137]). Hypsiprymnodontinae (p. [138]).
				Phalangeridae (p. [138])		Phalangerinae (p. [140]). Phascolarctinae (p. [142]). Phascolomyinae (p. [144]). Tarsipedinae (p. [145]).
				Epanorthidae (p. [145]).
		Polyprotodontia (p. [149])		Dasyuridae (p. [149]). Didelphyidae (p. [155]). Peramelidae (p. [156]). Notoryctidae (p. [158]).
EDENTATA (p. [161])		Xenarthra (p. [166])		Myrmecophagidae (p. [166]). Bradypodidae (p. [170]). Dasypodidae (p. [173]). Mylodontidae (p. [179]). Megalonychidae (p. [183]). Megatheriidae (p. [183]). Glyptodontidae (p. [184]).
		Nomarthra (p. [186])		Orycteropodidae (p. [187]). Manidae (p. [188]).
GANODONTA (p. [190])				Stylinodontidae (p. [191]). Conoryctidae (p. [193]).
UNGULATA (p. [195])		Condylarthra (p. [202]). Amblypoda (p. [205]). Ancylopoda (p. [211]). Typotheria (p. [212]). Toxodontia (p. [214]).
		Proboscidea (p. [216])		Elephantidae (p. [217]). Dinotheriidae (p. [231]).
		Hyracoidea (p. [232]).
		Perissodactyla (p. [235])		Equidae (p. [237]). Lophiodontidae (p. [247]). Palaeotheriidae (p. [247]). Tapiridae (p. [260]). Rhinocerotidae (p. [253]). Titanotheriidae (p. [264]).
		Litopterna (p. [267])		Macraucheniidae (p. [267]).
		Artiodactyla (p. [269])		Hippopotamidae (p. [273]). Suidae (p. [275]). Dicotylidae (p. [278]). Tragulidae (p. [282]). Proceratidae (p. [284]). Camelidae (p. [285]).
				Cervidae (p. [291])		Cervinae (p. [293]). Moschinae (p. [299]).
				Giraffidae (p. [301]). Antilocapridae (p. [306]). Bovidae (p. [307]). Anthracotheriidae (p. [328]). Caenotheriidae (p. [329]). Xiphodontidae (p. [329]). Oreodontidae (p. [330]). Anoplotheriidae (p. [332]).
SIRENIA (p. [333]).
CETACEA (p. [339])		Mystacoceti (p. [353])		Balaenopteridae (p. [355]). Balaenidae (p. [358]).
		Odontoceti (p. [362])		Physeteridae (p. [362])		Physeterinae (p. [363]). Ziphiinae (p. [367]).
				Delphinidae (p. [372]). Platanistidae (p. [380]). Squalodontidae (p. [384]).
		Archaeoceti (p. [384])		Zeuglodontidae (p. [384]).
CARNIVORA (p. [386])		Fissipedia (p. [387])		Felidae (p. [390]). Machaerodontidae (p. [401]).
				Viverridae (p. [403])		Euplerinae (p. [403]). Galidictiinae (p. [404]). Cryptoproctinae (p. [404]). Viverrinae (p. [405]). Herpestinae (p. [409]).
				Hyaenidae (p. [411]). Canidae (p. [413]). Procyonidae (p. [426]).
				Mustelidae (p. [431])		Melinae (p. [432]). Mustelinae (p. [433]). Lutrinae (p. [439]).
				Ursidae (p. [442]).
		Pinnipedia (p. [446])		Otariidae (p. [450]). Trichechidae (p. [451]). Phocidae (p. [452]).
CREODONTA (p. [455]).
RODENTIA (p. [458])		Simplicidentata (p. [462])		Anomaluridae (p. [462]). Soiuridae (p. [463]). Castoridae (p. [467]). Haplodontidae (p. [469]). Gliridae (p. [470]).
				Muridae (p. [471])		Murinae (p. [471]). Phlaeomyinae (p. [473]). Hydromyinae (p. [474]). Rhynchomyinae (p. [474]). Gerbillinae (p. [475]). Otomyinae (p. [475]). Dendromyinae (p. [476]). Lophiomyinae (p. [476]). Microtinae (p. [477]). Sigmodontinae (p. [479]). Neotominae (p. [480]).
				Bathyergidae (p. [480]). Spalacidae (p. [482]). Geomyidae (p. [483]). Heteromyidae (p. [484]). Dipodidae (p. [484]). Pedetidae (p. [486]).
				Octodontidae (p. [487])		Octodontinae (p. [487]). Loncherinae (p. [488]). Capromyinae (p. [489]).
				Ctenodactylidae (p. [490]). Caviidae (p. [491]). Dasyproctidae (p. [493]). Dinomyidae (p. [495]). Chinchillidae (p. [496]). Cercolabidae (p. [497]). Hystricidae (p. [499]).
		Duplicidentata (p. [502])		Leporidae (p. [502]). Lagomyidae (p. [505]).
TILLODONTIA (p. [506]).
INSECTIVORA (p. [508])		Insectivora Vera (p. [509])		Erinaceidae (p. [509]). Tupaiidae (p. [511]). Centetidae (p. [511]). Potamogalidae (p. [513]). Solenodontidae (p. [513]). Chrysochloridae (p. [514]). Macroscelidae (p. [515]). Talpidae (p. [516]). Soricidae (p. [518]).
		Dermoptera (p. [520])		Galeopithecidae (p. [520]).
CHIROPTERA (p. [521])		Megachiroptera (p. [524])		Pteropodidae (p. [524]).
		Microchiroptera (p. [526])		Rhinolophidae (p. [527]). Nycteridae (p. [527]). Vespertilionidae (p. [528]). Emballonuridae (p. [530]). Phyllostomatidae (p. [531]).
PRIMATES (p. [533])		Lemuroidea (p. [534])		Lemuridae (p. [538])		Indrisinae (p. [538]). Lemurinae (p. [540]). Galagininae (p. [542]). Lorisinae (p. [545]).
				Chiromyidae (p. [548]). Tarsiidae (p. [550]). Anaptomorphidae (p. [552]). Chriacidae (p. [552]). Megaladapididae (p. [554]).
		Anthropoidea (p. [554])		Hapalidae (p. [556]). Cebidae (p. [557]). Cercopithecidae (p. [562]). Simiidae (p. [570]). Hominidae (p. [585]).

CHAPTER I

INTRODUCTORY

The Mammalia form a group of vertebrated animals which roughly correspond with what are termed in popular language "quadrupeds," or with the still more vernacular terms of "beasts" or "animals." The name "Mammal" is derived from the most salient characteristic of the group, i.e. the possession of teats; but if the term were used in an absolutely strict etymological sense, it could not include the Monotremes, which, though they have mammary glands, have not fully-differentiated teats (see p. [16]). There are, however, as will be seen shortly, other characters which necessitate the inclusion of these egg-laying quadrupeds within the class Mammalia.

The Mammalia are unquestionably the highest of the Vertebrata. This statement, however, though generally acceptable, needs some explanation and justification. "Highest" implies perfection, or, at any rate, relative perfection. It might be said with perfect truth that a serpent is in its way an example of perfection of structure: not incommoded with limbs it can slip rapidly through the grass, swim like a fish, climb like a monkey, and dart upon its prey with rapidity and accuracy. It is an example of an extremely specialised reptile, the loss of the limbs being the most obvious way in which it is specialised from more generalised reptilian types. Specialisation in fact is often synonymous with degradation, and, this being the case, implies a restricted life. On the other hand, simplification is not always to be read as degeneration. The lower jaw, for instance, of mammals has fewer bones in it than that of reptiles, and is more concisely articulated to the skull; this implies greater efficiency

as a biting organ. The term highest, however, includes increased complexity as well as simplification, the two series of modifications being interwoven to form a more efficient organism. It cannot be doubted that the increased complexity of the brain of mammals raises them in the scale, as does also the complex and delicately adjusted series of bonelets which form the organ for the transmission of sound to the internal ear. The separation of the cavity containing the lungs, and the investment of the partition so formed with muscular fibres, renders the action of the lungs more effective; and there are other instances among the Mammalia of greater complexity of the various parts and organs of the body when compared with lower forms, which help to justify the term "highest" generally applied to these creatures.

Complexity and finish of structure are often accompanied by large size; and the Mammalia are, on the whole, larger than any other Vertebrates, and also contain the most colossal species. The huge Dinosaurs of the Mesozoic epoch, though among the largest of animals, are exceeded by the Whales; and the latter group includes the mightiest creature that exists or has ever existed, the eighty-five-feet-long Sibbald's Rorqual. Confining ourselves rigidly to facts, and avoiding all theorising on the possible relation between complexity and nicety of build and the capacity for increase in bulk, it is plain from the history of more than one group of mammals that increase in bulk accompanies specialisation of structure. The huge Dinocerata when compared with the ancestral Pantolambda teach us this, as do many similar examples. Within the mammalian group, as in the case of other Vertebrates, difference of size has a certain rough correspondence with difference of habitat. The Whales not only contain the largest of animals, but their average size is great; so too with the equally aquatic Sirenia and very aquatic Pinnipedia. Here the support offered by the water and the consequent decreased need for muscular power to neutralise the effects of gravity permit of an increase in bulk. Purely terrestrial animals come next; and finally arboreal, and, still more, "flying" mammals are of small size, since the maintenance of the position when moving and feeding needs enormous muscular effort.

The Mammals are more easily to be separated from the Vertebrates lying lower in the series than any of the latter are from each other in ascending order. A large number of

characters might be used in addition to those which will be made use of in the following brief catalogue of essential mammalian features, were it not for the low-placed Monotremata on the one hand and the highly specialised Whales on the other. Including those forms, the Mammalia are to be distinguished from all other Vertebrates by the following series of structural features, which will be expanded later into a short disquisition upon the general structure of the Mammalia. The class Mammalia may, in fact, be thus defined:—

Hair-clad Vertebrates, with cutaneous glands in the female, secreting milk for the nourishment of the young. Skull without prefrontal, postfrontal, quadrato-jugal, and some other bones, and with two occipital condyles formed entirely by the exoccipitals. Lower jaw composed of dentary bone only, articulating only with the squamosal. Ear bones a chain of three or four separate bonelets. Cervical vertebrae sharply distinguished from the dorsals, and if with free ribs, showing no transition between these and the thoracic ribs. Brain with four optic lobes. Lungs and heart separated from abdominal cavity by a muscular diaphragm. Heart with a single left aortic arch. Red blood-corpuscles non-nucleate.

The following characters are also very nearly universal, and in any case absolutely distinctive:—Cervical vertebrae, seven; vertebrae with epiphyses. Ankle-joint "cruro-tarsal," i.e. between the leg and the ankle, and not in the middle of the ankle.[^[1]] Attachment of the pelvis to the vertebral column pre-acetabular in position.

The Mammalia since they are hot-blooded creatures are more independent of temperature than reptiles; they are thus found spread over a wider area of the earth's surface. As however, though hot-blooded, they have not the powers of locomotion possessed by birds, they are not quite so widely distributed as are those animals. The Mammalia range up into the extreme north, but, excepting only forms mainly aquatic, such as the Sea Lions, are not known to occur on the Antarctic continent. With the exception of the flying Bats, indigenous mammals are totally absent from New Zealand; and it seems to be doubtful whether those supposed oceanic islands which have a mammalian fauna are really

oceanic in origin. The continents and oceans are peopled by rather over three thousand species of Mammalia, a number which is considerably less than that of either birds or reptiles. It seems clear that, so far at any rate as concerns the numbers of families and genera, the mammalian fauna of to-day is less varied than it was during the Mid-tertiary period, the heyday of mammalian life. It is rather remarkable to contrast in this way the mammals and the birds. The two classes of the animal kingdom seem to have come into being at about the same period; but the birds either have reached their culminating point to-day, or have not yet reached it. The Mammalia, on the other hand, multiplied to an extraordinary extent during the Eocene and the Miocene periods, and have since dwindled. The break is most marked at the close of the Pleistocene, and may be in part due to the direct influence of man. At present man exercises so enormous an effect, both directly and indirectly, that the future history of the Mammalia is probably foreshadowed by the instances of the White Rhinoceros and the Quagga. On the other hand, the economic usefulness of the Mammalia is greater than that of any other animals; and the next most important era in their history will be probably that of domesticity and "preservation."

CHAPTER II

STRUCTURE AND PRESENT DISTRIBUTION OF THE MAMMALIA

External Form.—It would be quite impossible for any one to confuse any other quadrupedal animal with a mammal. The body of a reptile is, as it were, slung between its limbs, like the body of an eighteenth century chariot between its four wheels; in the mammal the body is raised entirely above, and is supported by, the four limbs. The axes of these limbs too, as a general rule, are parallel with the vertical axis of the body of their possessor. There is thus a greater perfection of the relations of the limbs to the trunk from the point of view of a terrestrial creature, which has to use those limbs for rapid movement. The same perfection in these relations is to be seen, it should be observed, in such running forms among the lower Vertebrata as the Birds and the Dinosaurs, where the actual angulation of the limbs is as in the purely running Mammalia. These relations are of course absolutely lost in the aquatic Cetacea, and not marked in various burrowing creatures. The way in which the fore- and hind-limbs are angulated is considerably different in the two cases. In the latter, which are most used and, as it were, push on the anterior part of the body, the femur has its lower end directed forwards, the tibia and the fibula project backwards at the lower end, while the ankle and foot are again inclined in the same direction as the femur. With the fore-limbs there is not this regular alternation. The humerus is directed backwards, the fore-arm forwards, and the hand still more forwards. This angulation seems to facilitate movement, inasmuch as it is seen in even the Amphibia and the lower Reptiles, in which, however, the differences between the fore- and hind-limbs are less marked,

indicating therefore a less specialised condition of the limbs. It is an interesting fact that the angulation of the limbs is to some extent obliterated in very bulky creatures, and almost entirely so in the elephants (see p. [217]), which seem to need strong and straight pillars for the due support of their huge bodies.

The alertness and general intellectual superiority of mammals to all animals lying below them in the series (with the exception of the birds, which are in their way almost on a level with the Mammalia) are seen by their active and continuous movements. The lengthy periods of absolute motionlessness, so familiar to everybody in such a creature as the Crocodile, are unknown among the more typical Mammalia except indeed during sleep. This mental condition is clearly shown by the proportionate development of the external parts of all the organs of the higher senses. The Mammalia as a rule have well-developed, often extremely large, flaps of skin surrounding the entrance to the organ of hearing, often called "ears," but better termed "pinnae." These are provided with special muscles, and can be often moved and in many directions. The nose is always, or nearly always, very conspicuous by its naked character; by the large surface, often moist, which surrounds the nostrils; and again by the muscles, which enable this tract of the integument to be moved at will. The eyes, perhaps, are less marked in their predominance over the eyes of lower Vertebrates than are the ears and nose; but they are provided as a rule with upper and lower eyelids, as well as by a nictitating membrane as in lower Vertebrates. The apparent predominance of the senses of smell and hearing over that of sight appears to be marked in the Mammalia, and may account for their diversity of voice as well as of odour, and for the general sameness of coloration which distinguishes this group from the brilliantly-coloured birds and reptiles. The head, too, which bears these organs of special sense, is more obviously marked out from the neck and body than is the case with the duller creatures occupying the lower branches of the Vertebrate stem.

The Hair.—The Mammalia are absolutely distinguished from all other Vertebrates (or, for the matter of that, Invertebrates) by the possession of hair. To define a mammal as a Vertebrate with hair would be an entirely exclusive definition; even in the smooth Whales a few hairs at least are present, which may be

reduced to as few as two bristles on the lips. The term "hair," however, is apt to be somewhat loosely applied; it has been made use of to describe, for example, the slender processes of the chitinous skin of the Crustacea. It will be necessary, therefore, to enter into the microscopical structure and development of the mammalian hair. Hair is found in every mammal. The first appearance of a hair is a slight thickening of the stratum Malpighii of the epidermis, the cells taking part in this being

elongated and converging slightly above and below. Dr. Maurer has called attention to the remarkable likeness between the embryonic hair when at this stage and the simple sense-organs of lower Vertebrates. Later there is formed below this a denser aggregation of the corium, which ultimately becomes the papilla of the hair. This is the apparent homologue of the first formed part of a feather, which projects as a papilla before the epidermis has undergone any modification. Hence there is from the very first a difference between feathers and hairs—a difference which must be carefully borne in mind, especially when we consider the strong superficial resemblance between hairs and the simple barbless feathers. Still later the knob of epidermic cells becomes depressed into a tubular structure, which is lined with cells also derived from the stratum Malpighii, but is filled with a continuation of the more superficial cells of the epidermis. This is the hair-follicle, and from the epidermic cells arises the hair by direct metamorphosis of those cells; there is no excretion of the hair by the cells, but the cells become the hair. From the hair-follicle also grows out a pair of sebaceous glands, which serve to keep the fully-formed hair moist.

Fig. 2.—Four diagrams of stages in the development of a hair. A, Earliest stage in one of those mammals in which the dermal papilla appears first; B, C, D, three stages in the development of the hair in the human embryo. blb, Hair-bulb; crn, horny layer of the epidermis; foll, hair-follicle; grm, hair-germ; h, hair, in D, projecting on the surface; muc, Malpighian layer of epidermis; pp, dermal papilla; seb, developing sebaceous glands; sh.1, sh.2, inner and outer root-sheaths. (After Hertwig.)

Dr. Meijerle[^[2]] has lately described in some detail the

particular arrangement of the individual hairs among mammals; they are not by any manner of means scattered without order, but show a definite and regular arrangement, which varies with the animal. For instance, in an American Monkey (Midas), the hairs arise in threes—three hairs of equal size springing from the epidermis close together; in the Paca (Coelogenys) there are in each group three stout hairs alternating with three slender hairs. In some forms a number of hairs spring from a common point: in the Jerboa (Dipus) twelve or thirteen arise from a single hole; in Ursus arctos there is the same general plan, but there is one stout hair and four or five slender ones. There are numerous other complications and modifications, but the facts, although interesting, do not appear to throw any light upon the mutual affinities of the animals. Allied forms may have a very different arrangement, while in forms which have no near relationship the plan may be very similar, as is shown by the examples cited from Dr. Meijerle's paper. The groups of hairs, moreover, have themselves a definite placing, which the same anatomist has compared with the disposition of the bundles of hairs behind and between the scales of the Armadillo, and which has led him to the view that the ancestors of mammals were scaly creatures—a view also supported by Professor Max Weber,[^[3]]and not in itself unreasonable when we consider the numerous points of affinity between the primitive Mammalia and certain extinct forms of reptiles.[^[4]]

The hairs are greatly modified in form in different mammals and in different parts of their bodies. It is very commonly the case that a soft under-fur can be distinguished from the longer and coarser hairs, which to some extent hide the latter. Thus the "sealskin" of commerce is the under-fur of the Otaria ursina of the North. The coarser hairs may be further differentiated into bristles; these again into spines, such as those of the Hedgehog and of the Porcupine. Again, the flattening and agglutination of hairs seems to be responsible for the scales of the Manis

and for the horns of the Rhinoceros. It is a matter of common knowledge that upon the head of various animals, e.g. the Domestic Cat, long and sensitive hairs are developed, which are connected with the terminations of nerves, and perform a sensory, probably tactile function. These occur on the snout, above the eyes, and in the neighbourhood of the ears. It is an interesting fact that a tuft of quite similar hairs occurs on the hand of many mammals close to the wrist, which, at least in the case of Bassaricyon, are connected with a strong branch from the arm-nerve. These tufts also occur in Lemurs, in the Cat, various Rodents and Marsupials, and are probably quite general in mammals who "feel" with their fore-limbs;—in which, in fact, the fore-limbs are not exclusively running organs. That the last remaining hairs of the Cetacea are found upon the muzzle, is perhaps significant of the importance of these sensory bristles. The entire absence of hairs is quite common in this order, although traces of them are sometimes found in the embryo. The Sirenia, too, are comparatively hairless, as are also many Ungulates. Whether the presence of blubber in the former case and the existence of a very thick skin in the latter animals are facts which have had anything to do with the disappearance of hair or not, is a matter for further inquiry.

The intimate structure of the hair varies considerably. The variations concern the form of the hair, which may be round in transverse section, or so oval as to appear quite flat when the hair is examined in its entirety. The substance of the hair is made up of a central medulla or pith with a peripheral cortex; the latter is scaled, and the scales are often imbricated and with prominent edges. The amount of the two constituents also differs, and the cortex may be reduced to a series of bands surrounding only tracts of the enclosed pith. In the hair is contained the pigment to which the colour of mammals is chiefly due. Tracts of brightly-coloured skin may exist, as in the Apes of certain genera; but such structures are not general. The pigment of the hair seems to consist of those pigmentary substances known as melanins. It is remarkable to find such a uniform cause of coloration, when we consider the great variety of feather-pigments found in birds. The variations of colour of the hair of mammals are due to the unequal distribution of these brown pigments. There are very few mammals which can

be called brightly coloured. The Bats of the genus Kerivoula have been compared to large butterflies, and some of the Flying Squirrels have strongly-marked contrasts of reddish brown, white, and yellow. The same may be said of the spines of certain Porcupines. But we find in the hair no bright blues, greens, and reds such as are common among birds.

There are certain general facts about the coloration of mammals which require some notice here. Next to the usually sombre hues of these animals the general absence of secondary sexual coloration is noteworthy. In but a few cases among the Lemurs and Bats do we find any marked divergences in hues between males and females. Secondary sexual characters in mammals are, it is true, often exhibited by the great length of certain hair-tracts in the male, such as the mane of the Lion, the throat- and leg-tufts of the Barbary Sheep, and so forth; but apart from these, the secondary sexual characters of mammals are chiefly shown in size, e.g. the Gorilla, or in the presence of tusks, e.g. various Boars, or of horns, as in the Deer, etc. The coloration of mammals frequently exhibits conspicuous patterns of marking. These are in the form of longitudinal stripes, of cross-stripes, or of spots; the latter may be "solid" spots, or broken up, as in the Leopard and Jaguar, into groups of smaller spots arranged in a rosette-fashion. We never find in mammals the complicated "eyes" and other markings which occur in so many birds and in other lower Vertebrates. It is important to note that in the Mammalia whose sense of sight is quite keen there should be a practical absence of secondary sexual colours. As to the relationship of the various forms of marking that do occur, it seems clear that there has been a progression from a striped or spotted condition to uniform coloration. For we find that many Deer have spotted young; that the young Tapir of the New World is spotted, while its parents are uniform blackish brown; the strongly-marked spotting of the young Puma contrasts with the uniform brown of the adult; and the Lion cub, as every one knows, is also spotted, the adult lioness showing considerable traces of the spots.

The seasonal change in the colours of certain mammals is a subject upon which much has been written. The extreme of this is seen in those creatures, such as the Polar Hare and the Arctic Fox, which become entirely blanched in the winter, recovering

their darker coat in the spring. This is, however, only an extreme case of a change which is general. Most animals get a thicker fur in winter and exchange it for a lighter one in summer. And the hues of the coat change in correspondence.

Glands of the Skin.—The great variety of integumental glands possessed by the Mammalia distinguishes them from any group of lower Vertebrates. This variability, however, only concerns the anatomical structure of the glands in question. Histologically they are all of them apparently to be referred to one of two types, the sudoriparous or sweat gland and the sebaceous gland. Simple sweat and sebaceous glands are abundant in mammals, with but a few exceptions. The structures that we are now concerned with are agglomerations of these glands. The mammary glands will be treated of in connexion with the marsupium; they are either masses of sweat glands, or of sebaceous glands whose secretion has been converted into milk.

Many Carnivora possess glands opening to the exterior, near the anus, by a large orifice. These secrete various odoriferous substances, of which the well-known "civet" is an example. Other odoriferous glands are the musk glands of the Musk-deer and of the Beaver; the suborbital gland of many Antelopes; the dorsal gland of the Peccary, which has given the name of Dicotyles to the genus on account of its resemblance in form to a navel. This gland may be seen to secrete a clear watery fluid. The Elephant has a gland situated on the temple, which is said to secrete during certain periods only, and to be a warning to leave the animal alone. Very remarkable are the foot glands of certain species of Rhinoceros; they are not universally present in those animals, and are therefore useful as specific distinctions. On the back of the root of the tail in many Dogs are similar glands. The Gentle Lemur (Hapalemur) has a peculiar gland upon the arm, about the size of an almond, which in the male underlies a patch of spiny outgrowths. In Lemur varius is a hard patch of black skin which may be the remnants of such a gland. It is thought that the callosities on the legs of Horses and Asses are remnants of glands.

One of the most complex of these structures which has been examined microscopically exists in the Marsupial Myrmecobius.[^[5]] On the skin of the anterior part of the chest, just in front of the

sternum, is a naked patch of skin which is seen to be perforated by numerous pores. Besides the ordinary sebaceous and sweat glands there are a series of masses of glands, opening by larger orifices, which present the appearance of groups of sebaceous glands, and are of a racemose character; but the existence of muscular fibres in their coats seems to show that they should be referred rather to the sudoriparous series. Beneath the integument is a large compound tubular gland quite half an inch in diameter.

In Didelphys dimidiata there is a precisely similar glandular area and large underlying gland, the correspondence being remarkable in two Marsupials so distant in geographical position and affinities. Even among the Diprotodont genera there is something of the kind; for in Dorcopsis luctuosa and D. muelleri is a collection of four unusually large sebaceous follicles upon the throat, and in the Tree Kangaroo (Dendrolagus bennettii) there is the same collection of enlarged hair-follicles, though they are apparently somewhat reduced as compared with those of Dorcopsis. These are of course a few examples out of many.

It seems to be possible that the functions of these various glands is at least twofold. In the first place, they may serve, where predominant in one sex, to attract the sexes together. In the second place, the glands may be useful to enable a strayed animal of a gregarious species to regain the herd. It is perfectly conceivable too that in other cases the glands may be a protection, as they most undoubtedly are in the Skunk, from attacks. In connexion with the first, and more especially the second, of the possible uses of these glands, it is interesting to note that in purely terrestrial creatures, such as the Rhinoceros, the glands are situated on the feet, and would therefore taint the grass and herbage as the animal passed, and thus leave a track for the benefit of its mate. The same may be said of the rudimentary glands of Horses if they are really glands. The secretion of the "crumen" of Antelopes is sometimes deposited deliberately by Oreotragus upon surrounding objects, a proceeding which would attain the same end. One may even perhaps detect "mimicry" in the similar odours of certain animals. Prey may be lured to their destruction, or enemies frightened away. The defenceless Musk-deer may escape its foes by the suggestion of the musky odour of a crocodile. It is at any rate perfectly conceivable that the variety of odours among mammals may play a very

important part in their life, and it is perhaps worthy of note that birds with highly-variegated plumage are provided only with the uropygial gland, while mammals with usually dull and similar coloration have a great variety of skin glands. Scent is no doubt a sense of higher importance in mammals than in birds. The subject is one which will bear further study.

Nails and Claws.—Except for the Cetacea (where rudiments have been found in the foetus), the extremities of the fingers and of the toes of mammals are covered by, or encased in, horny epidermic plates, known as nails, claws, and hoofs.

The variety in the shape and development of these corneous sheaths to the digits is highly characteristic of mammals as opposed to lower Vertebrates. If we take extreme cases, such as the nail of the thumb in Man, the hoof of a Horse, and the claw of a Cat, it is easy to distinguish the three kinds of phalangeal horny coverings. But the differences become extinguished as we pass from these to related types. The nail of the little finger in Man approaches the claw-like form; and the hoofs of the Lama are almost claws in the sharpness of their extremities. On the whole it may be said that claws and hoofs embrace the bone which they cover, while nails lie only upon its dorsal surface. The form of the distal phalanx which bears the nail shows, however, two kinds of modification which do not support such a classification. When those phalanges are clad with hoofs or covered by a nail they end in a rounded and flattened termination. On the other hand, when they bear a claw they are themselves sharpened at the extremity and often grooved above.

The Marsupium.—It may appear to be unnecessary at this juncture to speak of the marsupial pouch, which is so usually believed to be a characteristic of the group Marsupialia. Rudiments of this structure have, however, been recently discovered in the higher mammals, and, as Dr. Klaatsch[^[6]] has remarked, all researches into the "history of the mammals culminate in the question whether the placental mammals pass through a marsupial stage or not." We cannot, therefore, look upon the marsupial pouch as a matter affecting only the Marsupials, though it is true that this organ is at present functional only in them and in the Monotremata.

Fig. 3.—Echidna hystrix. A, Lower surface of brooding female; B; dissection showing a dorsal view of the pouch and mammary glands; ††, the two tufts of hair in the lateral folds of the mammary pouch from which the secretion flows, b.m, Pouch; cl, cloaca; g.m, groups of mammary glands. (From Wiedersheim's Comparative Anatomy, after W. Haacke.)

In the Marsupials the pouch shelters the young, which are born in an exceedingly imperfect state, minute, nude, and blind, with a "larval" mouth fitted only to grasp in a permanent fashion the teat, upon which they are carefully fixed by the parent. But even later the pouch is made use of as a temporary harbour of refuge: from the pouch of female Kangaroos at the Zoological Gardens may frequently be observed to protrude the tail and hind-legs of a young Kangaroo as big as a Cat, and perfectly well able to take care of itself.

In the Monotremata (in Echidna) there is a deep fold of the skin which lodges the unhatched egg, and into which the mammary glands open, one on either side. This structure is only periodically developed, and arises from two rudiments, one corresponding to each mammary area; but in the female with eggs or young there is but a single deep depression, which occupies the same region of the body as the marsupial pouch of the

Marsupials.[^[7]] It is usually held that this structure is not of precisely the same morphological value as the pouch of the Marsupial; and the difference is expressed by terming the one (that of Echidna) the mammary pouch, and the other the marsupium. At first sight it may appear to be an unnecessary refinement to separate two structures which have so many and such obvious likenesses. It is not quite certain, however, that the difference is not even more profound than later opinions seem to indicate. The Monotremata not only have no teats, as has already been pointed out, but the mammary glands themselves are of a perfectly different nature to those of the higher mammals, including the Marsupials. There is therefore no a priori objection to the view that the accessory parts developed in connexion with the mammary glands should also be different. The teat of the higher Mammalia grows up round the area upon which the ducts of the mammary glands open; it is a fold of skin which eventually assumes the cylindrical form of the adult teat, and which includes the ducts of the milk glands. It has been suggested that the two folds of skin which form the mammary pouch of Echidna are to be looked upon as the equivalent of the commencing teat of the higher mammal.[^[8]] In this case it is clear that the marsupial folds of the Marsupial cannot correspond accurately with the apparently similar folds of Echidna, because there are teats as well. It is the teats which correspond to the marsupial folds of Echidna. This view is in apparent contradiction to an interesting discovery in a specimen of a Phalanger by Dr. Klaatsch.[^[9]] This Marsupial, like most others, has a well-developed marsupial pouch, in which the young are lodged at birth; but round two of the teats is another distinct fold on either side, the outer wall of which forms the general wall of the pouch. Dr. Klaatsch thinks that these smaller and included pouches are the equivalents of the mammary pouches of Echidna. They contain teats, but this comparison does not do away with the validity of Gegenbaur's suggestion already referred to, because the teats are (see above)

secondary. If this fact be fairly to be interpreted in the sense which Dr. Klaatsch attaches to it, we have an interesting case of the growth of a new organ out of and partly replacing an old organ. In the Monotremes there is a pouch which facilitates or performs both nutritive and protective functions; in the Phalanger these two functions are carried on in separate pouches; finally, in other Marsupials, there is a return to the undifferentiated state of affairs found in the Monotremata, but with the help of a new organ not found in them.

Fig. 4.—Diagram of the development of the nipple (in vertical section). A, Indifferent stage, glandular area flat; B, elevation of the glandular area with the nipple; C, elevation of the periphery of the glandular area into the false teat, a, Periphery of the glandular area; b, glandular area; gl, glands. (From Gegenbaur.)

Though so characteristic of Marsupials, the marsupial pouch is not always developed in them. It is present in all the Kangaroos, Wallabies, and Wombats, in fact in the Diprotodonts. It is also present in a number of the carnivorous Polyprotodont Marsupials; but in Phascologale it is only present in rudiment, and in Myrmecobius it is entirely obsolete. In the American Opossums the state of the pouch is variable. "Generally absent, sometimes merely composed of two lateral folds of skin separate at each end, rarely complete," is Mr. Thomas' summary in his definition of the family Didelphyidae.[^[10]] Another curious feature of the pouch in the Marsupials is the variability in the position of the mouth of the pouch: in all the Diprotodonts it looks forward; but in many Polyprotodonts it looks backward. This, however, has some connexion with the habitual attitude of the possessor: in the Kangaroo, leaping along on its hind-legs, it is requisite that the pouch should open forwards; but in the dog-like Thylacine, going on all fours, the fact that the pouch

opens backwards is less disadvantageous to the contained young.

The male Thylacine has a pouch which is quite or very nearly as well formed as in the female. There are also rudiments of a pouch in the male foetuses of many Marsupials, especially of those belonging to the Polyprotodont section of the order, though these rudiments are by no means confined to that subdivision. Up to so late a period as the age of four months (length 19.8 cm.) the male Dasyurus ursinus has a pouch.

We have now to consider the interesting series of facts relative to the permanence—in a rudimentary condition it is true—of the mammary pouch in the higher Mammalia, facts which seem to be an additional proof that they have been derived from an ancestor in which the pouch was an organ of functional importance. The first definite proof of the occurrence of a pouch in any mammal not a Marsupial or a Monotreme was made by Malkmus, who found this structure in a Sheep. It seems, however, that the structures found in the higher mammals are not always comparable to the marsupium of the Marsupials, but sometimes to the mammary pouch of the Monotreme. That the Marsupials are a side line, and not involved in the ancestry of the Eutheria, is an opinion which is at present widely held. At the same time it is reasonable to suppose that the original stock lying between the Prototheria and the Metatheria, whence the latter and the Eutheria have arisen, preserved both the mammary pouch of the lower mammal and the marsupium of the further-developed stage, as does Phalangista occasionally at the present day. Hence to find remnants of both structures in existing mammals would not so incredible. This is what Dr. Klaatsch believes to be the case. In certain Ungulates, including two species of Antelope, Dr. Klaatsch found very considerable rudiments of folds provided with unstriated muscular fibre; there were in the adult Cervicapra isabellina a pair of pouches, one on each side, and a rudiment of a second on either side; possibly this multiplication of the pouches has relation to the number of young. That there is more than one pouch makes a comparison with the mammary pouch rather than with the marsupium probable. The Ungulate teat, it must be remembered (see p. [16]), is a secondary teat; hence there is no difficulty in the comparison from this point of view. A pouch containing a primary

teat would of course be absolutely incomparable with a mammary pouch, because in that case the wall of the teat itself would be the pouch.

Mammals belonging to quite different Orders show traces more or less marked of a marsupium. In young Dogs the teats are borne upon an area where the skin is thinner, the covering of hair less dense than elsewhere—all points of resemblance to the inside of the pouch of a Marsupial; in addition to this there are traces of the sphincter marsupii muscle. In other Carnivora there are similar vestiges. In Lemur catta a more complete rudiment of a marsupial pouch is to be met with. In this Lemur the teats are both inguinal and pectoral; the skin in these regions is thin and but slightly hairy, and extends forwards as two bands of the same thinness and smoothness on each side of the densely hairy skin covering the sternum. This area is sharply separated from the rest of the integument by a fold which runs parallel to the longitudinal axis of the body, and can be comparable with nothing save the rudiment of the marsupial fold.

One is tempted to wonder how far the habit which certain Lemurs have of carrying their young across the abdomen with the tail wrapped round the body of the mother is a reminiscence of a marsupial pouch.

Skeleton.

The skeleton of the Mammalia consists almost solely of the endoskeleton. It is only among the Edentata that an exoskeleton of bony plates in the skin is met with. As in other Vertebrates, the skeleton is divisible into an axial portion, the skull and vertebral column, and an appendicular skeleton, that of the limbs. The bones of mammals are well ossified, and in the adult there are but few and small tracts of cartilage left.

Vertebral Column.—The vertebral column of the mammals, like that of the higher Vertebrata, consists of a number of separate and fully-ossified vertebrae.

The constitution of a vertebra upon which all the usual processes are marked is as follows:—There is first of all the body or centrum of the vertebra, a massive piece of bone shaped like a disc or a cylinder. The centra of contiguous vertebrae

are separated by a certain amount of fibrous tissue forming the intervertebral disc, and the apposed surfaces of the centra are as a rule nearly flat. In this last feature, and in the important fact that the centra are ossified from three distinct centres, the anterior and posterior pieces ("epiphyses") remaining distinct for a time, even for a long time (as in the Whales), the centra in the mammals differ from those of reptiles and birds. The epiphyses are not found throughout the vertebral column of the lowly-organised Monotremata, and they do not appear to exist in the Sirenia.

From each side of the centrum on the dorsal side arises a process of bone which meets its fellow in the middle line above, and is from there often prolonged into a spine. A canal is thus formed which lodges the spinal cord. This arch of bone is known as the neural arch, and the dorsal process of the same as the spinous process. The sides of the neural arch bear oval facets, by which successive vertebrae articulate with one another: those situated anteriorly are the anterior zygapophyses, while those on the posterior aspect of the arch are the posterior zygapophyses; these articular facets do not exist in the tail-region of many mammals, e.g. Whales.

In addition to the dorsal median spinous process of the

vertebra there may be a ventral median process, arising of course from the centrum, termed the hypapophysis.

From the sides of the neural arch, or from the centrum itself, there is commonly a longer or shorter process on each side, known as the transverse process. This is sometimes formed of two distinct processes, one above the other; in such cases the upper part is called a diapophysis, the lower a parapophysis.

The neural arch may also bear other lateral processes, of which one directed forwards is the metapophysis, the other directed backwards the anapophysis.

The series of bones which constitute the vertebral column can be divided into regions. It is possible to recognise cervical, dorsal, lumbar, sacral, and caudal vertebrae. In the case of animals with only rudimentary hind-limbs, such as the Whales, there is no recognisable sacral region. The neck or cervical vertebrae are nearly always seven in number. The well-known exceptions are the Manatee, where there are six, and certain Sloths, where there are six, eight, or nine. These rare exceptions only accentuate the very remarkable constancy in number, which is very distinctive of the mammals as compared with lower Vertebrata. There are of course abnormalities, the last cervical, and sometimes the last two, assuming the characters of the ensuing dorsals, by developing a more or less complete rib. There are also recorded examples of Bradypus, in which the number of cervicals is increased to ten. The characteristics, then, of the cervical vertebrae are, in the first place, that they do not normally bear free ribs, and that there is a break as a rule between the last cervical and the first dorsal on this account. In birds, for example, the cervicals, differing in number in different families and genera, gradually approach the dorsals by the gradually lengthening ribs. The transverse processes of the vertebrae are commonly perforated by a canal for the vertebral artery, and are bifid at their extremities. In some Ungulates these vertebrae, moreover, approximate to the vertebrae of lower Vertebrata in the fact that there are ball and socket joints between the centra, instead of only the fibrous discs of the remaining vertebrae.

The first two vertebrae of the series are always very different from those which follow. The first is termed the

atlas, and articulates with the skull. The most remarkable fact about this bone (shared, however, by lower Vertebrates) is that its centrum is detached from it and attached to the next vertebra, in connexion with which it will be referred to immediately. The whole bone thus gets a ring-like form, and the salient processes of other vertebrae are but little developed, with the exception of the transverse processes, which are wide and wing-like. In many Marsupials, such as the Wombat and Kangaroo, the arch of the atlas is open below, there being no centre of ossification. In others, such as Thylacinus, there is a distinct nodule of bone in this situation not concrescent with the rest of the arch.

Fig. 9.—Atlas of Kangaroo.... (From Parker and Haswell's Zoology.)

The second vertebra, which is known as the axis or epistropheus, is a compound structure, the anterior "odontoid process," which fits into the ring of the atlas, being in reality the detached centrum of that vertebra.[^[11]] It is a curious fact about that process that it has independently become spoon-shaped in two divisions of Ungulates; that it has become so seems to be shown by the fact that in the earlier types of both it has the simple peg-like form, which is the prevailing form. The cervical

vertebrae are occasionally wholly (Right Whales) or partially (many Whales, Jerboa, certain Edentates) welded into a combined mass. Indications of this have even been recorded in the human subject.

The dorsal vertebrae vary greatly in number: nine (Hyperoodon) seems to be the lowest number existing normally; while there may be as many as nineteen, as in Centetes, or twenty-two, as in Hyrax. These vertebrae are to be defined by the fact that they carry ribs, and the first one or two lumbars are often "converted into" dorsals by the appearance of a small supernumerary rib. The spinous processes of these vertebrae are commonly long, and sometimes very long. It is only among the Glyptodons that any of these vertebrae are fused together into a mass.

The lumbar vertebrae, which follow the dorsal, vary greatly in number. There are as few as two in the whale Neobalaena, as many as seventeen in Tursiops; this group, the Cetacea, contains the extremes. Nine lumbars are found in the Lemurs Indris and Loris. As a rule the number of lumbars is to some extent dependent upon that of the dorsals. It often happens that the number of thoraco-lumbar vertebrae is constant for a given group. Thus the Artiodactyles have nineteen of these vertebrae, and the Perissodactyles as a rule twenty-three. A greater number of dorsals implies a smaller number of lumbars, and of course vice versa. The existence of a sacral region formed of a

number of vertebrae fused together and supported by the pelvic girdle is characteristic of the mammals, but is not found in the Cetacea and the Sirenia, where functional hind-limbs are wanting. Strictly speaking, the sacrum is limited to the two or three vertebrae whose expanded transverse processes meet the ilia. But to these are or may be added a variable number of vertebrae withdrawn from both the lumbar and the caudal series, which unite with each other to form the massive piece of bone which constitutes the sacrum of the adult.

The caudal vertebrae complete the series. They begin in as fully developed a condition as the lumbars, with well-marked transverse processes, etc.; but they end as no more than centra, from which sometimes tiny outgrowths represent in a rudimentary way the neural arches, etc. Very often the caudal vertebrae are furnished with ventral, generally

-shaped, appendages, the chevron bones or intercentra.[^[12]] These are

particularly conspicuous in the Whales and in the Edentates. In the former group the occurrence of the first intercentrum serves to mark the separation of the caudal from the lumbar series. The number of caudals varies from three in Man—and those quite rudimentary—to nearly fifty in Manis macrura and Microgale longicaudata.

Fig. 14.—Lateral view of skull of a Dog. C.occ, Occipital condyle; F, frontal; F.inf, infra-orbital foramen; Jg, jugal; Jm, premaxilla; L, lachrymal; M, maxilla; Maud, external auditory meatus; Md, mandible; N, nasal; P, parietal; Pal, palatine; Pjt, process of squamosal; Pt, pterygoid; Sph, alisphenoid; Sq, squamosal; Sq.occ, supraoccipital; T, tympanic. (From Wiedersheim's Comparative Anatomy.)

The Skull.—The skull in the Mammalia differs from that of the lower Vertebrata in a number of important features, which will be enumerated in the following brief sketch of its structure. In the first place, the skull is a more consolidated whole than in reptiles; the number of elements entering into its formation is less, and they are on the whole more firmly welded together than in Vertebrates standing below the Mammalia in the series. Thus in the cranial region the post- and pre-frontals, the post-orbitals and the supra-orbitals have disappeared, though now and again we are reminded of their occurrence in the ancestors of the Mammalia by a separate ossification corresponding to some of the bones. Nowhere is this consolidation seen with greater clearness than in the lower jaw. That bone, or rather each half of it, is in mammals formed of one bone, the dentary (to which occasionally, as it appears, a separate mento-Meckelian

ossification may be added). The angular, splenial, and all the other elements of the reptilian jaw have vanished, though the numerous points from which the mammalian dentary ossifies is a reminiscence of a former state of affairs; and here again an occasional continuance of the separation is preserved, as the case observed by Professor Albrecht of a separate supra-angular bone in a Rorqual attests. Among other reptilian bones that are not to be found in the mammalian skull are the basipterygoids, quadrato-jugal, and supratemporal. A few of these bones, however, though no longer traceable in the adult skull save in cases of what we term abnormalities, do find their representatives in the foetal skull. Professor Parker, for example, has described a supra-orbital in the embryo Hedgehog; a supratemporal also appears to be occasionally independent.

Fig. 15.—Head of a Human embryo of the fourth month. Dissected to show the auditory ossicles, tympanic ring, and Meckel's cartilage, with the hyoid and thyroid apparatus. All these parts are delineated on a larger scale than the rest of the skull. an, Tympanic ring; b.hy, basihyal element; hy, so-called hyoid bone; in, incus; md, bony mandible; ml, malleus; st, stapes; tp, tympanum; tr, trachea; I. (mk), first skeletal (mandibular) arch (Meckel's cartilage); II. second skeletal (hyoid) arch; III. third (first branchial) arch; IV. V. fourth and fifth arches (thyroid cartilage). (From Wiedersheim's Structure of Man.)

In the mode of the articulation of the lower jaw to the skull the Mammalia apparently, perhaps really, differ from other Vertebrates. In the Amphibia and Reptilia, with which groups alone any comparisons are profitable, the lower jaw articulates by means of a quadrate bone, which may be movably or firmly attached to the skull. In the mammals the articulation of the lower jaw is with the squamosal. The nature of this articulation is one of the most debated points in comparative anatomy. Seeing that Professor Kingsley[^[13]] in the most recent contribution to the subject quotes no less than fifty-two different views, many of which are more or less convergent, it will be obvious that in a work like the present the matter cannot be treated exhaustively. As, however, Professor Kingsley justly says that "no single bone occupies a more important position in the discussion of the origin of the Mammalia than does the quadrate," and with equal justice adds that "upon the answer given as to its fate in this group depends, in large measure, the broader problem of the phylogeny of the Mammalia," it becomes, or indeed has long been, a matter which cannot be ignored in any work dealing with the mammals. A simple view, due to the late Dr. Baur and to Professor Dollo, commends itself at first sight as meeting the case. The last-named author holds, or held, that in all the higher Vertebrates it is at least on a priori grounds likely that two such characteristically vertebrate features as the lower jaw and the chain of bones bringing the outer world

into communication with the internal organ of hearing would be homologous throughout the series. He believed, therefore, that the entire chain of ossicula auditus in the mammal is equal to the columella of the reptile, since their relations are the same to the tympanum on the one hand and to the foramen ovale on the other; and that the lower jaw articulates in the same way in both. It follows, therefore, that the glenoid part of the squamosal must be the quadrate which has become ankylosed with it after the fashion of concentration in the mammalian skull that has already been referred to. The fact that occasionally the glenoid part of the squamosal is a separate bone[^[14]] appeared to confirm this way of looking at the

matter. But the hall-mark of truth is not always simplicity; indeed the converse appears to be frequently the case. And on the whole this view does not commend itself to zoologists at present. For it must be borne in mind that the lower jaw of the mammal is not the precise equivalent of that of the reptiles. Apart from the membrane bones, which may be collectively the equivalents of the dentary of the mammal, there is the cartilaginous articular bone to be considered, which forms the connexion between the rest of the jaw and the quadrate in reptiles. Even in the Anomodontia, whose relations to the Mammalia are considered elsewhere, there is this bone. But in these reptiles the articular bone articulates not only with the quadrate, but also to a large extent with the squamosal, the quadrate shrinking in size and developing processes which give to it very much the look of either the incus or the malleus of the mammalian ear. In fact it seems on the whole to fit in with the views of the majority, as well as with a fair interpretation of the facts of embryology, to consider that the chain of ear bones in the mammal is not the equivalent of the columella of the reptile, but that the stapes of the mammal is the columella, and that the articulare is represented by the malleus and the quadrate by the incus. It is very interesting to note this entire change of function in the bones in question. Bones which in the reptile serve as a means of attachment of the lower jaw to the skull are used in the mammal to convey the waves of sound from the tympanum of the ear to the internal organ of hearing.

Another important and diagnostic feature in the mammalian skull is that the first vertebra of the vertebral column always articulates with two separate occipital condyles, which are borne by the exoccipital bones and formed mainly though not entirely by them. Certain Anomodontia form the nearest approach to the mammals in this particular. The two condyles of Amphibia are purely exoccipital in origin.

In the Mammalia, unlike what is found in lower Vertebrates (but here again the Anomodontia form at least a partial exception), the jugal arch does not connect the face with the quadrate, for, as already said, that bone does not exist, in the Sauropsidan form, in mammals. This arch passes from the squamosal to the maxillary, and has but one separate bone in addition to those two, viz. the jugal or malar.

Fig. 16.—Under surface of the cranium of a Dog. × ½. apf, Anterior palatine foramen; as, posterior opening of alisphenoid canal; AS, alisphenoid; BO, basioccipital; BS, basisphenoid; cf, condylar foramen; eam, external auditory meatus; Ex.O, exoccipital; flm, foramen lacerum medium; flp, foramen lacerum posterius; fm, foramen magnum; fo, foramen ovale; fr, foramen rotundum; Fr, frontal; gf, glenoid fossa; gp, post-glenoid process; Ma, malar; Mx, maxilla; oc, occipital condyle; op, optic foramen; Per, mastoid portion of periotic; pgf, post-glenoid fossa; Pl, palatine; PMx, premaxilla; pp, paroccipital process; ppf, posterior palatine foramen; PS, presphenoid; Pt, pterygoid; sf, sphenoidal fissure or foramen lacerum anterius; sm, stylomastoid foramen; SO, supraoccipital; Sq, zygomatic process of squamosal; Ty, tympanic bulla; Vo, vomer. (From Flower's Osteology.)

In connexion with the elaboration of the chain of auditory ossicles it is very usual for mammals to possess a thin inflated bone, sometimes partly or entirely formed out of the tympanic bone, and known as the tympanic bulla. Whether this structure is thin and inflated or thick and depressed in form it is characteristic of the mammals, and does not occur below them in the series. But it is not present in all mammals. It is absent, for example, in the Monotremes. When it is present it is sometimes formed from other bones, as, for instance, from the alisphenoids. The tympanic ring has been held to be the equivalent of the quadrate. It is more probably the quadrato-jugal.[^[15]]

Fig. 17.—A, First thoracic skeletal segment for comparison with B, fifth cervical vertebra (Man), b.v. Body of vertebra; c, first thoracic rib; c′, cervical rib (which has become united with the transverse process, tr), the two enclosing the costo-transverse foramen (f.c.t); st, sternum; zy, articular process of the arch (zygapophysis). (From Wiedersheim's Structure of Man.)

Ribs.—All mammals are furnished with ribs, of which the number of pairs differs considerably from group to group, or it may be even from species to species. The ribs are attached as a rule by two heads, of which one, the capitulum, arises as a rule between two centra of successive vertebrae. The other, the tuberculum, springs from the transverse process. Only in the Monotremes

are there ribs with but one, the capitular, head. In the posterior part of the series the two heads often gradually coalesce, so that there comes to be but one, the capitular, head. The Whales also, at least the Whalebone Whales, are exceptional in possessing but one head to the ribs, which is the capitular. The first rib joins the sternum below, and a variable number after this have the same attachment. There are always a number of ribs, sometimes called floating ribs, which have no sternal attachment. In the Whalebone Whales it is the first rib alone which is so attached. As a rule, to which the Whales mentioned are again an exception, the rib is divided into at least two regions—the vertebral portion which is always ossified, and the sternal moiety which is usually cartilaginous. This is, however, often very short in the first rib. They are, however, ossified in the Armadillos and in some other animals. Between the vertebral and sternal portions an intermediate tract is separated off and ossified in the Monotremata. The ribs of existing mammals belong only to the dorsal region of the vertebral column, but there are traces of lumbar ribs and also of cervical ribs. In the Monotremata, indeed, these latter

are persistently free for a very long period, and in some cases never become ankylosed with their vertebrae. But it should be noted that in this group there is no approximation to the state of affairs which exists in many lower Vertebrates, where there is a gradual transition between the ribs of the cervical and those of the dorsal region of the vertebral column; for that of the seventh ribs in Monotremes is smaller than those which precede it.

The Sternum.—All the Mammalia so far as is known possess a sternum. This is the bone, or series of bones (sternebrae), which lies upon the ventral surface of the chest, and to which the ribs are attached below. The development of the sternum has been shown to take place from the fusion of the ribs below into two lateral bands, one on each side; the approximation of these bands forms the single and unpaired sternum of most mammals. Very considerable traces, however, of the paired state of the sternal bones often exist; thus in the Sperm Whale the first piece of the sternum is divided into two by a longitudinal division, and the second piece is longitudinally grooved. The development of the sternum out of the fused ends of ribs is shown in a more complete condition in some species of Manis than in many other mammals. Thus in M. tricuspis the last ribs of those which are attached to the sternum are completely fused together into a single piece on each side.[^[16]] As a general rule the last ribs which come into relation with the sternum do so only in an imperfect way, being simply firmly attached at their sides to, but not fused with, the last ribs which are definitely articulated with the sternum. Contrary to what is found in lower

Vertebrates, the sternum of the Mammalia consists of a series of pieces, as many as eight or nine or even sixteen in Choloepus, of which the first is called the manubrium sterni, and the last the ensiform cartilage, xiphisternum, or xiphoid process. The latter often remains largely cartilaginous throughout life; in fact this is generally but not universally the case with that part of the breastbone. The most extraordinary modification of the xiphoid process is seen in the African species of the genus Manis, where it diverges into two long cartilages, which run back to the pelvis and then, curving round, run forwards and fuse together in the middle line anteriorly. These processes serve for the attachment of certain tongue-muscles. They were looked upon by Professor Parker as the equivalents of the "abdominal ribs" of reptiles elsewhere non-existent among mammals. This view is not, however, usually held. The manubrium sterni is often keeled in the middle line below; this is so with the Bats, which thus approach the birds, and probably for the same reason, i.e. the need of an enlarged origin for the pectoral muscle, which is concerned in the movements of flight. In many forms this part of the sternum is much broader than the pieces which follow; this is so with the Viscacha. In the Pig the precise reverse is seen, the manubrium being narrower than the rest of the sternal bonelets. It will be noticed, however, that in this and similar cases there are no clavicles. Ribs are attached between the successive pieces of the sternum. When the sternum is reduced, as it is in the Cetacea and in the Sirenia, it is the intermediate part of the series of bones which becomes abbreviated or vanishes. The Sperm Whale has only a manubrium sterni and a following piece belonging to the mesosternum. It is fair to say that the xiphoid process and the rest of the sternum have disappeared, since among the Toothed Whales a progressive shortening of the sternum can be seen. In the Whalebone Whales the sternum is still further reduced; the manubrium is alone left, and to it are attached but a single pair of ribs. In Balaena, however, a rudimentary

piece, apparently comparable to a xiphoid process, has been detected.

From the instances which have been described, as well as from the mode of development of the sternum and from the number of free ribs, i.e. ribs which are not attached to it, it would seem that the sternum has undergone a considerable reduction in its size. This reduction may be possibly accounted for by the need for respiratory activity, which is clearly increased by a less-marked fixity of the walls of the thoracic cavity. In the case of the Whales one can hardly help coming to that conclusion. The arrangement in the Monotremata does not, however, point in the same direction; for these animals are precisely like the higher Mammalia in the reduction of the sternum and of the number of ribs which reach it.

Fig. 22.—Shoulder girdle of Ornithorhynchus. c¹, c², c³, First, second, third ribs; cl, clavicle; e.c, epicoracoid; es′ and es″, interclavicle (episternum); m.c, metacoracoid; m.s, manubrium sterni; sc, scapula; st, sternebra. (From Wiedersheim's Structure of Man.)

Fig. 23.—Episternum of an embryo Mole. (After A. Götte.) cl, Clavicle; es′, central portion of the episternum; es″, lateral portion of the same; r.c, costal ribs; st, sternum. (The figure was constructed from two consecutive horizontal sections.) (From Wiedersheim's Structure of Man.)

The Episternum.—The Mammalia are as a rule to be distinguished from lower Vertebrates by the absence of an episternum, or interclavicle as it is also called. In the Monotremata, however, there is a large

-shaped bone which does not overlie the sternum as in reptiles, but is anterior to it. The relations of this bone to the clavicles seem to leave no doubt that it is the equivalent of the Lacertilian interclavicle or episternum. The Monotremata are not, however, the only mammals in which this structure is to be seen. The Mole in the embryonic condition is

provided with pieces of bone which overlie the manubrium sterni and are attached to the clavicles, and are no doubt to be regarded as the same structure. Probably in many mammals the manubrium will be found to be partly made up of corresponding rudiments. In any case, vestiges of an episternum in the shape of two minute ossicles have been discovered in Man, lying in front of the manubrium. They have been termed ossa suprasternalia. In Man and in the Mole the paired nature of the episternum is clearly apparent. It has been suggested that this structure in its entirety belongs to the clavicles, just as the sternum belongs to the ribs; i.e. that it formed out of the approximated and fused ends of the clavicles. Dr. Mivart[^[17]] figured a good many years since a pair of ossicles in Mycetes, lying in one case between the ends of the clavicles and the manubrium sterni, and in another example anterior to the ventral ends of the clavicles. Gegenbaur has figured a

pair of similar bones in the Hamster.[^[18]] It is possible that these are to be referred to the same category. It has also been suggested that these supposed episternal rudiments are the vestiges of a pair of cervical ribs.

Fig. 24.—Episternal vestiges in Man. cl, Clavicle, sawn through; es, "episternum" (sternoclavicular cartilage); l′, interclavicular ligament; l″, costoclavicular ligament; m.s, manubrium sterni; o.s, ossa suprasternalia; r.c, first rib; st, sternum. (From Wiedersheim's Structure of Man.)

The Pectoral Girdle.—The skeleton by which the fore-limb is connected with the trunk is known as the Pectoral Girdle. The main part of this girdle is formed by the large scapula, or blade-bone as it is often termed. The coracoidal elements will be dealt with later. The scapula is not firmly connected with the backbone; it is attached merely by muscles, thus presenting a great difference from the corresponding pelvic girdle. The reason for this difference is not easy to understand. On the one hand it may be pointed out that in all running animals at any rate there is a greater need for the fixation in a particularly firm way of the hind-limbs; but, again, in the climbing creatures both limbs would, one might suppose, be bettered by a firm fixation. It must be remembered, however, that in the latter case the same result is at least partly brought about by a well-developed clavicle, which fixes the girdle to the sternum and so to the vertebral column by means of the ribs.

Broadly speaking, too, the fore-limbs require a greater freedom and variety of movement than the hind-limbs, which are supports

for or serve to push along the rapidly-moving body. Stronger fixation is therefore a greater necessity posteriorly than anteriorly. In any case, whatever the explanation, this important difference exists.

The shoulder-blade of mammals is as a rule a much-flattened bone with a ridge on the outer surface known as the spine; this ridge ends in a freely-projecting process, the acromion, from which a branch often arises known as the metacromion. This gives a bifurcate appearance to the end of the ridge. The spine is less developed and the scapula is narrower in such animals as the Dog and the Deer which simply run, and whose fore-limbs therefore are not endowed with the complexity of movement seen, for instance, in the Apes.

Fig. 27.—Right scapula of Dolphin (Tursiops tursio). × ¼. a, Acromion; af, prescapular fossa; c, coracoid; gc, glenoid cavity; pf, postscapular fossa. (From Flower's Osteology.)

Fig. 28.—Side view of right half of shoulder girdle of a young Echidna (Echidna hystrix). × ⅔. a, Acromion; c, coracoid; cb, coracoid border; cl, clavicle; css, coraco-scapular suture; ec, epicoracoid; gb, glenoid border; gc, glenoid cavity; ic, interclavicle; pf, postscapular fossa; ps, presternum; s, spine; ss, suprascapular epiphysis; ssf, subscapular fossa. (From Flower's Osteology.)

It has been pointed out that the area which lies in front of the spine, the prescapular lamina, is most extensively developed in such animals as perform complex movements with the fore-limbs. The Sea Lion and the Great Anteater are cited by Professor G. B. Howes as examples of this preponderance of the anterior portion of the scapula over that which lies behind the spine. The general shape of the scapula varies considerably among the different orders of mammals; but it always presents the characters mentioned, which are nowhere seen among the Sauropsida except among certain Anomodonts, which will be duly referred to (see p. [90]). The most conspicuous divergences from the normal are to be found in the Cetacea and the Monotremata. In the former the acromion is approximated so nearly to the anterior border of the blade-bone that the prescapular fossa is reduced to a very small area; and in Platanista the acromion actually coincides with the anterior border, so that that fossa actually disappears. In the Whales, too, the scapula is as a rule very broad, especially above; it has frequently a fan-like contour. In the Monotremata the acromion also coincides with the anterior border of the scapula; but the sameness of appearance which it thus presents (in this feature) to the Cetacean scapula is

apparently not due to real resemblance. What has happened in the Monotremata is, that the prescapular fossa is so enormously expanded that it occupies the whole of the inner side of the blade-bone, while the subscapular fossa which, so to speak, should occupy that situation, has been thus pushed round to the front, where it is divided from the postscapular fossa by a slight ridge only.

The clavicle is a bone which varies much in mammals. It is sometimes indeed, as in the Ungulata, entirely absent; in other forms it shows varying degrees of retrocession in importance; it is only in climbing, burrowing, digging, and flying mammals that it is really well developed.

Fig. 29.—Shoulder girdle, with upper end of sternum (inner surface) of Shrew (Sorex), after Parker, × 7. a, Acromion; c, coracoid; cl, clavicle; ec, partially ossified "epicoracoid" of Parker, or rudiment of the sternal extremity of the coracoid; ''ma'', metacromial process; mss, ossified "mesoscapular segment"; ost, omosternum; pc, rudiment of precoracoid (Parker); ps, presternum; sr¹, first sternal rib; sr², second sternal rib. (From Flower's Osteology.)

In the higher Mammalia the coracoid[^[19]] is present, but does not reach the sternum as in the Monotremata. It is known to human anatomists as the coracoid process of the scapula. It has been found, however, by Professor Howes[^[20]] and others, that this process really consists of two separate centres of ossification, forming two separate bonelets, which in the adult become firmly ankylosed to each other and to the scapula. These two separate bones have been met with in the embryo of Lepus, Sciurus, and the young of various other mammals belonging to very diverse orders, such as Edentates and Primates. The separation even occasionally persists in the adult. The question is, What is the relation of these bonelets to the coracoid of the Monotremata and to the corresponding regions of reptiles? Professor Howes terms the lower patch of bone the metacoracoid and the upper the epicoracoid;

the former is alone concerned with the glenoid cavity. It must therefore, one would suppose, correspond to the "coracoid" of the Monotremata, while the upper piece of bone is the epicoracoid process of that mammal. The Mammalia, therefore, higher as well as lower, differ from the reptiles in that the coracoid is formed of two bones, the exceptions being, among some other extinct forms, certain of the Anomodontia, a group which it will be recollected is the nearest of all reptiles to the mammals.

Fig. 30.—Distal extremity of the humerus to show Epicondylar Foramina. A, In Hatteria; B, in a Lizard (Lacerta ocellata); C, in the Domestic Cat; D, in Man. c.e, External condyle; c.i, internal condyle. In A the two foramina are developed (at i, the entepicondylar; at ii, the ectepicondylar). The only canal (†) present in the Lizard (B) is on the external ulnar side, in the cartilaginous distal extremity. In Man (D) an entepicondylar process (pr) is sometimes developed and continued as a fibrous band. (From Wiedersheim's Anatomy of Man.)

The Fore-limb.—The humerus is of varying length among mammals. A feature which it sometimes shares with the humerus of lower forms is the presence of an entepicondylar foramen, a defect of ossification situated above the inner condyle of that bone which transmits a nerve. The same foramen and an additional ectepicondylar foramen are found in the ancient reptilian type Hatteria (Sphenodon); it occurs also in the Anomodont reptiles. It is as a rule only the lower forms among mammals which show this foramen; thus it is present in the Mole and absent in the

Horse. The fact that it is occasionally met with in Man is an additional proof of the, in many respects, ancient structure of the highest type of Primate.

The radius and the ulna, which together constitute the fore-arm, are both present in a large number of mammals, but the ulna tends to vanish in the purely walking and digitigrade Ungulates, being present, however, in the more ancient forms of these Ungulates. In Man and in many other mammals the radius can be moved from its normal position and crossed over the ulna; this movement of pronation has been permanently fixed in the Elephant, where the bones are crossed but cannot be altered in position by the contractions of any muscles. Other types agree with the Elephant in this fixation of the two bones.

Fig. 31.—Bones of fore-arm and manus of Mole (Talpa europaea). × 2. c, Cuneiform; ce, centrale; l, lunar; m, magnum; p, pisiform; R, radius; rs, radial sesamoid (falciform); s, scaphoid; td, trapezoid; tm, trapezium; U, ulna; u, unciform; I-V, the digits. (From Flower's Osteology.)

The bones of the wrist show great variation among mammals. The greatest number present are to be seen in such a type as the Mole. Here we have a proximal row, consisting of the scaphoid, lunar, cuneiform, and pisiform, which are arranged in their proper order, beginning with that on the radial side of the limb, that side which bears the first digit. A second row articulates proximally with these bonelets and distally with the metacarpals; the bones composing it are, mentioning them in the same order, trapezium, trapezoid, centrale, magnum, unciform.

The centrale does not, however, really belong to the distal carpal row, and is as a rule situated in the middle of the carpus away from articulation with the metacarpals. It is a bone which is not commonly present in the mammalian hand, but is present in various lower forms, such as the Beaver and Hyrax. It also occurs in such high types as the majority of Monkeys; it is to be found in the Human foetal carpus. Many extinct forms possessed a separate centrale. Its importance in the formation of the interlocking condition of the Ungulate foot is referred to later,

on p. [196]. The only mammal which appears to have the proper five bones in the distal row of the carpus corresponding to the five metacarpals is Hyperoodon, where this state of affairs at least occasionally occurs. The final bone of that series, the unciform, seems to represent two bones fused. Very often the carpus is reduced by the fusion of certain of the carpal bones; thus among the Carnivora it is usual for the scaphoid and the lunar to be fused. It is interestingly significant that these bones retain their distinctness in the ancestral Creodonts. In many Ungulates the trapezium vanishes. The reduction of the toes in fact implies a reduction of the separate elements of the carpus.

As to the digits of the mammalian hand, the greatest number is five, the various supplementary bonelets known as prepollex and postminimus being, it is now generally held, merely supplementary ossifications not representing the rudiments of pre-existing fingers. They may, however, bear claws.[^[21]] The number of phalanges which follow upon the metacarpals is almost constantly three in the mammals, excepting for the thumb, which has only two. This is highly characteristic of the group as opposed to reptiles and birds, and the increase in the number of these bones in the Whales and to a very faint degree in the Sirenia is a special reduplication, which will be mentioned when those animals are treated of.

The Pelvic Girdle.—The pelvic girdle or hip girdle is the combined set of bones which are attached on the one hand to the sacrum and on the other articulate with the hind-limb. Four distinct elements are to be recognised in each "os innominatum," the name given to the conjoined bones of each half of the entire pelvis. These are:—the ilium, which articulates with the sacrum; the ischium, which is posterior; the pubis, which is anterior; and finally, a small element, the cotyloid, which lies within the acetabular cavity where the femur articulates. The epipubes of the Monotreme and the Marsupial are dealt with elsewhere (see p. [116]) as they are peculiar to those groups.

Professor Huxley pointed out many years since that while the Eutherian Mammalia differ from the reptiles in the fact that the axis of the ilium lies at a less angle with that of the sacrum,

Ornithorhynchus comes nearest to the reptile in the fact that this axis is nearly at right angles to that of the sacrum. It is particularly interesting to find that this peculiarity of Ornithorhynchus is only acquired later in life, and that the pelvis of the foetus conforms in these angles to the adults of other mammalian groups. In any case, the backward rotation of the pelvis is a mammalian characteristic, and it is most nearly approached among reptiles by the extinct Anomodontia, whose affinities to mammals will be dealt with on a later page (p. [90]). Another peculiarity of the mammalian pelvis appears to be the cotyloid bone already referred to. In the Rabbit this bone completely shuts out the pubis from any share in the acetabular cavity; later it ankyloses with that bone. In Ornithorhynchus the cotyloid or os acetabuli is a larger element of the girdle than is the pubis. In other mammals, therefore, it seems to be a rudimentary structure. But it seems to be a bone peculiar to and thus distinctive of the mammals as compared with other vertebrates. The acetabular cavity is perforated in Echidna as in birds; but in certain Rodents the same region is very thin and only closed by membrane, as in Circolabes villosus.

The number and the arrangement of the bones in the hind-limb correspond exactly to those of the fore-limb. The femur, which corresponds to the humerus, shows some diversities of form. The neck, which follows upon the almost globular head, the surface of articulation to the acetabular cavity of the pelvis, has two roughened areas or tuberosities for the insertions of muscles. A third such area, known as the third trochanter, is present or absent as the case may be, and its presence or absence is of systematic import. As a general rule the thigh-bones of the ancient types of mammals are smoother and less roughened by the presence of these three trochanters than in their modern representatives. The radius and the ulna are represented in the hind-leg by the tibia and the fibula. These bones are not crossed, and do not allow of rotation as is the case with the radius and the ulna. In Ungulate animals there is the same tendency to the shortening and rudimentary character of the fibula that occurs in the case of the ulna, but it is more marked. It has been shown in tracing the history of fossil Ungulates that the hind-limbs in their degree of degeneration are as a rule ahead of the fore-limbs. This is natural when we reflect that

the hind-limbs must have preceded the fore-limbs in their thorough adaptation to the cursorial mode of progression. In the Mammalia the ankle-joint is always what is termed cruro-tarsal, i.e. between the ends of the limb-bones and the proximal row of tarsals; not in the middle of the tarsus as in some Sauropsida (reptiles and birds). The bones of the ankle are much like those of the hand; but there are never more than two bones in the proximal row, which are the astragalus and the calcaneum. The former is perhaps to be looked upon as the equivalent of the cuneiform and lunar together. But the views as to the homologies of the tarsal bones differ widely. Below these is the navicular, regarded as a centrale. The distal row of the tarsus has four bones, three cuneiforms and a cuboid. Reduction is effected by the soldering together of two cuneiforms as in the Horse, by the fusion of the navicular and cuboid as in the Deer. No mammal has more than five toes, and the number tends to become reduced in cursorial animals (Rodents, Ungulates, Kangaroos).

Fig. 32.—Anterior aspect of right femur of Rhinoceros (Rhinoceros indicus). × ½. h, Head; t, great trochanter; t′, third trochanter. (From Flower's Osteology.)

Teeth.—The teeth of the Mammalia[^[22]] differ from those of other vertebrated animals in a number of important points. These, however, entirely concern the form of the adult teeth, their position in the mouth, and the succession of the series of teeth. Developmentally and histologically there are no fundamental divergences from the teeth of vertebrates lower in the scale.

In mammals, as for example in the Dog, the teeth consist of three kinds of tissue—the enamel, the dentine, and the cement. The enamel is derived from the epidermis of the mouth cavity, and the two remaining constituents from the underlying dermis. The teeth originate quite independently of the jaws, with which they are later so intimately connected; the independence of origin being one of the facts upon which the current theory

of the nature of teeth is founded. It has been pointed out that the scales of the Elasmobranch fishes consist of a cap of enamel upon a base of dentine, the former being derived from the epidermis and modelled upon a papilla of the dermis whose cells secrete the dentine. The fact that similar structures arise within the mouth (i.e. the teeth) is explicable when it is remembered that the mouth itself is a late invagination from the outside of the body, and that therefore the retention by its tissues of the capacity to produce such structures is not remarkable.

Fig. 33.—Diagrammatic sections of various forms of teeth. I, Incisor or tusk of Elephant, with pulp cavity persistently open at base; II, Human incisor during development, with root imperfectly formed, and pulp cavity widely open at base; III, completely formed Human incisor, with pulp cavity opening by a contracted aperture at base of root; IV, Human molar with broad crown and two roots; V, molar of the Ox, with the enamel covering the crown deeply folded, and the depressions filled up with cement; the surface is worn by use, otherwise the enamel coating would be continuous at the top of the ridges. In all the figures the enamel is black, the pulp white; the dentine represented by horizontal lines, and the cement by dots. (After Flower and Lydekker.)

The relations of the three constituents of the tooth in its simplest form is shown in the accompanying diagram, where the intimate structure of the enamel, dentine, and cement (or crusta petrosa as it is sometimes called) is not indicated. The latter has the closest resemblance to bone. The dentine is traversed by fine canals which run parallel to each other and anastomose here and there. The enamel is formed of long prismatic fibres, and is excessively hard in structure, containing less animal matter than the other tooth tissues. To this fact is frequently

due the complicated patterns upon the grinding teeth of Ungulates, which are produced by the wearing away of the dentine and the cement, and the resistance of the enamel.

The centre of the tooth papilla remains soft and forms the pulp of the tooth, which is continuous with the underlying tissues of the gum by a fine canal or a wide cavity as the case may be. In teeth which persistently grow throughout the lifetime of the animal, as for example the incisors of the Rodents, there is a wide intercommunication between the cavity of the tooth and the tissues of the gum; only a narrow canal exists in, for instance, the teeth of Man, and in fact in the vast majority of cases. The three constituents of the typical teeth are not, however, found in all mammals; the layer which is sometimes wanting is the enamel. This is the case with most Edentates; but the interesting discovery has been made (by Tomes) that in the Armadillo there is a downgrowth of the epidermis similar to that which forms the enamel in other mammals, a rudimentary "enamel organ."

Teeth are present in nearly all the Mammalia; and where they are absent there is frequently some evidence to show that the loss is a recent one. The Whalebone Whales, the Monotremata, Manis, and the American Anteaters among the Edentata are devoid of teeth in the adult state. In several of these instances, however, more or less rudimentary teeth have been found, which either never cut the gums or else become lost early in life. The latter is the case with Ornithorhynchus, where there are teeth up to maturity (see p. [113]). Kükenthal has found germs of teeth in Whales, and Röse in the Oriental Manis. The loss of the teeth in these cases seems to have some relation to the nature of the food. In ant-eating mammals, as in the Anteaters and Echidna, the ants are licked up by the long and viscid tongue, and require no mastication. Yet it must be remembered that Orycteropus is also an anteater, like the Marsupial Myrmecobius, both of which genera have teeth.

The first of the essential peculiarities of the mammalian teeth as compared with those of other vertebrates concerns the position of the teeth in the mouth. There is no undoubted mammal extinct or living in which the teeth are attached to any bones other than the dentary, the maxilla, and the

premaxilla. There are no vomerine, palatine, or pterygoid teeth, such as are met with in Amphibia and Reptilia.

The other peculiarities of the mammalian teeth, though true of the great majority of cases, are none of them absolutely universal.

But it is necessary to go into the subject at some length on account of the great importance which has been laid upon the teeth in deciding questions of relationship; moreover, largely no doubt on account of their hardness and imperishability, our knowledge of certain extinct forms of Mammalia is entirely based upon a few scattered teeth; while of some others, notably of the Triassic and Jurassic genera, there is not a great deal of evidence except that which is furnished by the teeth. Indeed the important place which odontography holds in comparative anatomy is from many points of view to be regretted, though inevitable. "In hardly any other system of organs of vertebrated animals," remarks Dr. Leche, "is there so much danger of confounding the results of convergence of development with true homologies, for scarcely any other set of organs is less conservative and more completely subservient to the lightest impulse from without." Affinities as indicated by the teeth are sometimes in direct contradiction to those afforded by other organs; or, as in the case of the simple Toothed Whales, no evidence of any kind is forthcoming. Dr. Leche has pointed out that, judged merely from its teeth, Arctictis would be referred to the Raccoons, though it is really a Viverrid; while Bassariscus, which Sir W. Flower showed to be a Raccoon, is in its teeth a Viverrid. Mr. Bateson has been obliged to hamper the subject with another difficulty.

In dealing with the variations of teeth,[^[23]] Mr. Bateson has brought together an immense number of facts, which tend to prove that the variability of these structures is much greater than had been previously recognised; that this variability is often symmetrical; and that in some animals, as in "Canis cancrivorus, a South American fox, the majority showed some abnormality." When we learn from Mr. Bateson that "of Felis fontanieri, an aberrant leopard, two skulls only are known, both showing dental abnormalities," it seems dangerous to rear too lofty a superstructure upon a single fossil jaw. It must be noted too that,

contrary to the prevailing superstition, it is not domestic animals which show the greatest amount of tooth variation. As to special homologies between tooth and tooth, with which we shall deal on a later page, Mr. Bateson has urged almost insuperable difficulties.

Fig. 34.—Skull of Dasyurus (lateral view). al.sph, Alisphenoid; ang, angular process of mandible; fr, frontal; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; oc.cond, occipital condyle; par, parietal; par.oc, paroccipital process; p.max, premaxilla; s.oc, supraoccipital; sq, squamosal; sq′, zygomatic process of squamosal. (From Parker and Haswell's Zoology.)

Fig. 35.—Upper and lower teeth of one side of the mouth of a Dolphin (Lagenorhynchus), illustrating the homodont type of dentition in a mammal. (After Flower and Lydekker.)

The teeth of the Mammalia are almost without exception "heterodont," i.e. they show differences of structure in different parts of the mouth. As a general rule, teeth can be grouped into cutting incisors, sharp conical canines, and molars, with a surface which is in the majority of cases suited for grinding. In this they contrast with the majority of the lower vertebrates, where the teeth are "homodont" (or, better, homoeodont), i.e. all more or less similar and not fitted by change of form to perform different duties. But there are exceptions on both sides. In

the Toothed Whales the teeth are homodont, as they are in the frog and in most reptiles; on the other hand, some of the remarkable reptiles belonging to Professor Huxley's order of the Anomodontia have distinct canines, and show other differentiations in their teeth.

A second characteristic of the mammalian dentition is the limited number of the teeth, which rarely exceeds fifty-four. Here again the Toothed Whales are an exception, the number of their teeth being as great as in many reptiles. In the Mammalia the number of the teeth is fixed (excepting of course for abnormalities), while in reptiles there is frequently no precise normal. Two regions may be distinguished in every tooth—the crown and the root; the latter, as its name denotes, is imbedded in the gum, while the crown is the freely-projecting summit of the tooth. The varying proportions of these two regions of the tooth enables us to divide teeth into two series—the brachyodont and the hypselodont; in the latter the crown is developed at the expense of the root, which is small; the hypselodont tooth is one that grows from a persistent pulp or, at any rate, one that is long open. Brachyodont teeth on the contrary have narrow canals running into the dentine. The primitive form of the tooth seems undoubtedly to be a conical single-rooted tooth, such as is now preserved in the Toothed Whales and in the canine teeth of nearly all animals. The development of the teeth, that is, the simple bell-shaped form of the enamel organ, seems to go some way towards proving this; but it is quite another question whether we can fairly regard the Whales as having retained this early form of tooth. In their case the simplification, as is so often the case where organs are simplified, seems to be rather degeneration than retention of primitive characters. But this is a matter which must be deferred for the present.

The incisor teeth are generally of simple structure and nearly always single rooted. In the Rodents, in the extinct Tillodontia and in Diprotodont Marsupials, they have grown large, and, as has been already stated, they increase in size continuously from the growing pulp. These teeth have a layer of enamel only on the anterior face, which keeps a sharp chisel-like edge upon them by reason of the fact that the harder enamel is worn away more slowly than the comparatively soft dentine. The

"horn" of the Narwhal is another modification of an incisor, as are the tusks of Elephants. Among the Lemurs the incisors are denticulate, and serve to clean the fur in a comb-like fashion. This is markedly the case in Galeopithecus. The incisors are sometimes totally absent, as in the Sloths, sometimes partially absent, as in many Artiodactyles, where the lower incisors bite against a callous pad in the upper jaw, in which no trace of incisors has been found.

Canine teeth are present in the majority of mammals, but are absent without a single exception from the jaws of the Rodentia. The canine tooth of the upper jaw is that tooth which comes immediately after the suture dividing the premaxillary from the maxillary bone. The canines are as a rule simple conical teeth, with but a single root; indeed they resemble what we may presume to have been the first kind of tooth developed in mammals. In this they resemble also as a general rule the foregoing incisors. But instances are known where the canines are implanted by two roots. This is to be seen in Triconodon, in the pig Hyotherium, in the Mole and some other Insectivores, and in Galeopithecus, where the incisors also may be thus implanted in the jaw. Furthermore, the simple condition of the crown of the tooth may be departed from. This is the case with a Fruit Bat belonging to the genus Pteralopex. In the more primitive Mammalia it is common to find no great difference between the canines and incisors; such is the case with the early Ungulate types of Eocene times, such as Xiphodon. In modern mammals, however, especially among the Carnivora, the canines tend to become larger and stronger than the incisors, and in some of the Cats and in the Walrus these teeth are represented by enormous offensive tusks. It is not rare for the canines of male animals to be larger than those of their mates. There are also cases such as the Musk-deer and the Kanchil where the male alone possesses these teeth, but only in the upper jaw. The teeth which follow the canines are known as the grinders or cheek teeth, or more technically as premolars and molars. These two latter terms separate teeth which arise at different periods, and their use will be explained later. In the meantime it may be pointed out that the cheek teeth are the teeth which show the greatest amount of variation in their structure; this is shown by the number and variety of the cusps in which

the biting surface ends. The grinding teeth vary from simple one-cusped teeth, precisely like canines, to teeth with an enormous number of separate tubercles. In the former case it is hard to distinguish between incisors, canines, and cheek teeth in the lower jaw, where no suture separates the bone. Moreover it is quite common for the first cheek tooth in the lower jaw to have the characters of a canine, while the true canine approximates in its form to the antecedent incisors. This is so, for instance, with the Lemurs, where the first premolar is caniniform, and the canine shares in the curious procumbent attitude which distinguishes the lower incisors of many of those animals.

A variable number of the anterior cheek teeth may be little more than simple conical teeth; but the rest of the set are commonly more complicated. No definite laws can be laid down as to the complication of the posterior as compared with the anterior set. Broadly speaking, it is purely herbivorous creatures in which the least difference can be detected at the two extremities, and which are at the same time the most elaborately decorated with tubercles and ridges. The converse is true that in purely carnivorous animals, including insect- and fish-eating forms, there is the greatest difference between the anterior set of grinding teeth and those which follow. In these two respects such animals as a Lemur and a Rhinoceros occupy the extremes. Furthermore, it may be said that omnivorous creatures lie, as their diet would suggest, in an intermediate position. Generally speaking, when there is a marked difference between the first premolar and molars at the end of the series, there is a gradual approximation in structure of a progressive kind. The tubercles become more numerous in successive teeth; but the corollary which is apparently deducible from this, i.e. that the last molar is the most elaborate of the series, is by no means always true. The last cheek tooth indeed is often degenerate. On the other hand, it is very markedly the largest of the series in such diverse types as the Elephant, the hog Phacochoerus, and the Rodent Hydrochoerus. It is a rule that the cheek teeth of the upper jaw are more complicated than the corresponding teeth of the lower jaw.

The structure of the cheek teeth is very diverse among the Mammalia. Broadly, two types are to be recognised. There are

teeth in which the grinding surface is raised into a series of two, to many, tubercles sharper or blunter as the case may be;—sharper and fewer at the same time in carnivorous and especially in insectivorous types, more abundant in omnivorous animals. To this form of tooth the term "bunodont" is applied. There is no doubt that this is the earliest type of tooth; but whether the fewer or the more cusped condition is the primitive one is a question that is reserved for consideration at the end of the present chapter. The other type of grinding tooth is known as "lophodont." This is exemplified by such types as the Perissodactyla and Ungulates generally, and by the Rodents. The tooth is traversed by ridges which have generally a transverse direction to the long axis of the jaw in which the tooth lies. The ridges may be regarded as having been developed between tubercles which they connect and whose distinctness as tubercles is thereby destroyed. Lophodont teeth are only found in vegetable-feeding animals.

Fig. 36.—Molar teeth of Aceratherium platycephalum. × ½. m.1-m.3., Molars; mh, metaloph; p.1-p.4, premolars; ph, protoloph; ps.f, parastyle fossa; te, tetartocone. (After Osborn.)

The special characteristics of the teeth of various groups of animals will be considered further under the accounts of the several orders of recent and fossil Mammalia.

Fig. 37.—Two stages in the development of the teeth of a Mammal (diagrammatic sections). alv, Bone of alveolus; dent, dentine; dent.s, dental sac; en, enamel; en.m, enamel membrane; en.m², enamel membrane of permanent tooth; en.plp, enamel pulp; gr, dental groove; lam, dental lamina; lam′, part of dental lamina which grows downwards below the tooth germ; n, neck connecting germs of milk and permanent tooth; pap, dental papilla; pap², dental papilla of permanent tooth. (After O. Hertwig.)

A very general feature of the teeth of the Mammalia is what is usually termed the diphyodont dentition. In the majority of cases there are two sets of teeth developed, of which the first lasts for a comparatively short time, and is termed on account of its usual time of appearance the "milk dentition"; this is replaced later by the permanent dentition. In lower vertebrates the teeth are replaced as worn away. There is not, however, so great an antithesis in this matter between the Mammalia

and other vertebrates as was at one time assumed. But in order to explain this very important part of the subject it will be necessary to give some account of the development of the teeth. The type selected is the Hedgehog, which has been recently and carefully described by Dr. Leche of Stockholm, which type has furthermore the advantage of being a "central" type of mammal. The first step in the formation of the teeth is a continuous invagination of the epithelium covering the jaw to form a deepish wall of tissue running in the thickness of the jaw; this is perfectly continuous from end to end of the lower jaw. From this "common enamel germ" (Schmelzleiste of the Germans[^[24]]) "special enamel germs" (Schmelzorgane, enamel organs) are developed here and there as thickenings in the form of buds

which arise on the outer side of the fold of epithelium and some way above its lower termination. These ultimately acquire a bell-like form, and are as it were moulded on to a thickened concentration of the dermis beneath; they then become separate from the downgrowth of the epithelium whence they have arisen. Finally, each of the eight germs becomes one of the milk teeth of the animal. The lower end of the sheet of invaginated epithelium, the common enamel germ, is the seat of the formation of the second set of teeth, of which, however, in the animal under consideration, there are only two in each jaw. But corresponding to each of the enamel germs of the milk dentition, with the exception of the first two molars, there is a slight thickening of the end of the common enamel germ, which at a certain stage is indistinguishable from the thickening which will become one of the permanent teeth. We have thus the diphyodont arrangement. But this does not exhaust the series of rudimentary teeth, though no more come to maturity than those whose development has already been touched upon. In the upper jaw a small outgrowth of the common enamel germ arises above and to the outer side of the enamel germ of the third milk incisor; this does not develop any further, but its resemblance to the commencing germ of a tooth seems to indicate that it is the remnant of a tooth series antecedent to the milk series. Furthermore, there are indications in the fourth premolar of a fourth series of teeth posterior in appearance to the permanent dentition. We arrive therefore at the important conclusion that although here as elsewhere there are only two sets of calcified teeth ever developed, there are feeble though unmistakable remains of two other series, one antecedent to and the other posterior to the diphyodont dentition. The gap therefore which separates the mammalian dentition from that of reptiles is less than has hitherto appeared. Dr. Leche also carefully studied the tooth development of Iguana; he found that in this lizard there are four series of teeth which come to maturity, and a rudimentary series antecedent to these which never produces fully formed teeth.

In a few mammals there is a kind of dentition known as the monophyodont, in which only one series of teeth reaches maturity; where in fact there is no replacement of a milk series by a permanent dentition. Of the monophyodont dentition Whales form an example. The Marsupials are very nearly an instance of the

same phenomenon; for Sir W. Flower showed, and Mr. Thomas confirmed his discovery, that only one tooth, according to Mr. Thomas the fourth premolar, is replaced in that group. But even the purely monophyodont dentition of the Toothed Whales is a more apparent than real contrast to the diphyodont dentition elsewhere prevalent. An investigation of the embryos of various Toothed Whales by Dr. Kükenthal and by Dr. Leche has brought to light the highly important fact that two dentitions are present, but that one only comes to maturity; from this fact obviously follows the interesting question:—To which of the two dentitions of more normal Mammalia does the monophyodont dentition of the Whales and Marsupials belong? To this question a clear answer is fortunately possible. As has been pointed out in the foregoing sketch of tooth development, and has been illustrated in the figures, the milk teeth develop as lateral outgrowths of the common enamel germ, while the permanent teeth arise from the end of the same band of tissue. This fact enables it to be stated apparently beyond a doubt that in the Whales and in the Marsupials it is the milk dentition which is the only one to arrive at maturity. Thus the earlier theoretical conclusion that the Marsupial dentition "is a secondary dentition with only one tooth of the primary set left," is proved on embryological grounds to be untrue. But there are other monophyodont animals than those already mentioned.[^[25]] Orycteropus, the Cape Anteater, is an example. Mr. Thomas has lately discovered that in this Edentate there is a set of minute though calcified milk teeth which probably never cut the gum; here we have a different sort of monophyodontism, in which the teeth belong to the second and not to the first set. Between the latter condition and the diphyodont state are intermediate stages. Thus in the Sea Lions the milk teeth are developed but disappear early, probably before the animal is born.

In the typical diphyodont dentition, such as is exhibited for example in Man and the vast majority of mammals, the milk teeth eventually completely disappear and are entirely replaced by the permanent set of teeth, with the exception, of course, of the molars, which though they are developed late belong to the milk series.

Their correspondence with the milk series is shown in an interesting way by the close resemblance which the last milk premolar often bears to the first molar. These two extremes of dentition, i.e. purely monophyodont and, excepting for the molars, purely diphyodont, are however connected by an intermediate state of affairs, which is represented by more than one stage. In Borhyaena (probably a Sparassodont) the incisors and the canines and two out of the four premolars belong to the permanent dentition, while the two remaining premolars and of course the three molars are of the milk series. Prothylacinus, a genus belonging to the same group, has a dentition which is a step or two further advanced in the direction of the recent Marsupials. We find, according to Ameghino,[^[26]] whose conclusions are accepted by Mr. Lydekker, that the incisors, canines, and two premolars belong to the milk series, while the permanent series is represented only by the two remaining premolars. We can tabulate this series as follows:—

(1) Purely monophyodont, with teeth only of the first set—Toothed Whales.

(2) Incompletely monophyodont, as in the Marsupials, where there is a milk dentition with only one tooth replaced.[^[27]]

(3) Incompletely diphyodont, with the dentition made up partly of milk, partly of permanent teeth, as in Borhyaena.

(4) Diphyodont, where all the teeth except the molars are of the second set; this characterises nearly all the mammals.

As we pass from older forms to their more recent representatives there is as a rule a progressive development of the form of the teeth. This is especially marked among the Ungulata. The extremely complicated type of tooth found in such a form as the existing Horse can be traced back through a series of stages to a tooth in which the crown is marked by a few separated tubercles or cusps. Arrived at this point, the differences between the teeth of ancestral Horses and ancestral Rhinoceroses and Tapirs are hard to distinguish with accuracy; and the same difficulty is experienced in attempting to give a definition of other large orders by the characters of the teeth, such as will apply to the Eocene or

even earlier representatives of these families. Fig. 36 (p. [51]) illustrating a series of mammalian teeth will illustrate the above remarks. That there is such a convergence in tooth structure shows that it is, theoretically at least, possible to determine the ancestral form of the mammalian tooth. Practically, however, the difficulties which beset such theorising are great; that there are such divergent and such strongly-held antithetical views is sufficient proof of this. Two main views hold the field: one, which has found most favour in America, and is due chiefly to the labours and persuasiveness of Professors Cope, Scott, Osborn, and others, is known as "trituberculy."[^[28]] The alternative view, as urged by Forsyth Major, Woodward, and Goodrich, attempts to show that the dentition of the original mammal included grinding teeth which were multicuspidate or "multitubercular." There is much to be said for both views, and something to be said against both.

Fig. 38.—Molar teeth of A, Phenacodus, and B, the Creodont Palaeonictis. End, endoconid; hld, hypoconulid; hyd, hypoconid; med, metaconid; prd, protoconid. (After Osborn and Wortman.)

This question is, however, wrapped up in a wider one. Its solution depends upon the ancestry of mammals. If the Mammalia are to be derived from reptiles with simple conical teeth, then the first stage in the development of trituberculy is proved. On the other hand, however, the evidence is gradually growing that the Theromorpha represent more nearly than any non-mammalian group with which we are acquainted the probable ancestral form of the mammals. These animals offer some support to both the leading views. Cynognathus had triconodont teeth which, as will be pointed out later, are a theoretically intermediate stage in the evolution of tritubercular teeth; on the other hand, the teeth of Diademodon and some others are multituberculate, and have been very properly compared to the multitubercular teeth of such primitive mammalia as the Ornithorhynchus. Professor Osborn is no doubt correct in italicising a remark of an anonymous writer in Science to the effect that in Diademodon the teeth, though multitubercular, show the prevalence of three cusps arranged in the tritubercular fashion.

But this may be only a proof that the multitubercular antedates the tritubercular. It may be, indeed, that the mammalian tooth was already differentiated among the mammal-like Saurians and that from such a form as Cynognathus the Eutheria and other forms in which a tritubercular arrangement can be detected were evolved, and from such form as Tritylodon the Monotrematous branch of the mammals. This way of looking at the matter harmonises a much-disputed question, but involves a diphyletic origin of the mammals—an origin which for other reasons is not without its supporters.

We shall now attempt to give a general idea of the facts and arguments which support or tend to support "trituberculy." As a matter of fact the name is inaccurate; for the holders of this view do not derive the mammalian molar from a trituberculate condition, but in the first place from a simple cone such as that of a crocodile!

To this main and at first only cusp came as a reinforcement an additional cusp at each side, or rather at each end, having regard to their position with reference to the long axis of the jaw. This stage is the "triconodont" stage, and teeth exist among living as well as extinct mammals which show this early form of tooth. We have, indeed, the genus Triconodon, so named on that very account. Among living mammals the Seals and the Thylacine all show some triconodont teeth. A Toothed Whale, it may be remarked, is a living example of a mammal with monoconodont teeth. The three primary cusps, as the supporters of Cope's theory of trituberculism denominate them, are termed respectively the protocone, paracone, and metacone, or, if they are in the teeth of the lower jaw, protoconid, paraconid, and metaconid. At a slightly later stage, or coincidently, a rim partly surrounded the crown of the tooth; the rim is known as the cingulum, and from a prominent elevation of this rim a fourth cusp, the hypocone, was developed. The three main cones then moved, or rather two of them moved, so as to form a triangle; this is the tritubercular stage. Teeth of this pattern are common, and occur in such ancient forms as Insectivora and Lemurs, besides numerous extinct groups. An amendment has been suggested, and that is to term the teeth with the simple primitive triangle "trigonodont," and to reserve the term tritubercular for those teeth in which the hypocone has appeared. The platform bearing the hypocone widened into the

"talon"; and this ledge became produced into two additional cusps, the hypoconule or hypoconulid, and the ectocone or ectoconid. Thus the typical sextuberculate tooth of the primitive Ungulate, and indeed of many primitive Eutherians, is arrived at. From this the still further complicated teeth of modern Ungulates can be derived by further additions or fusions, etc.[^[29]] On the other hand, the development of the Primate molar stops short at the stage of four cusps.

Fig. 39.—Epitome of the evolution of a cusped tooth. 1, Reptile; 2, Dromatherium; 3, Microconodon; 4, Spalacotherium; me, metaconid; pa, paraconid; pr, protoconid; 5, Amphitherium. (After Osborn.)

That such a series can be traced is an undoubted fact. Every stage exists, or has existed. But whether the stages can be connected or not is quite another question. It is by three main lines of argument that the view here sketched out in brief is supported. In the first place, the tracing of the pedigrees of many groups of mammals has met with very considerable success; and it is clear that as we pass from the living Horse and Rhinoceros, with their complicated molars, to their forerunners, we find that both can be referred to a primitive Ungulate molar with but six cusps. Going still further back to the lowest Eocene and ancestral type as it appears, Euprotogonia, we still find in the molar tooth the sextubercular plan of structure. We can hardly get further back in the evolution of the Perissodactyles with any probability of security. On the other hand, many facts point to a fundamental relationship between the primitive Ungulates and the early Creodonts. The latter frequently show plainly tritubercular molars. Such Ungulates as Euprotogonia and Protogonodon, though sex- or quinque-tubercular as to their molars, have a distinctly prevailing trituberculism, when the size and importance of three of the cusps is taken into account. But this

lacks finality as a convincing proof of the tritubercular tooth as a primitive Ungulate tooth.

Professor Osborn has ingeniously utilised certain deviations from the normal type of tooth structure (for the group) in favour of his strongly-urged opinions. If the stages of development have been as he suggests, a retrogression would naturally be in the inverse order; thus the "apparently 'triconodont' lower molar of Thylacinus" may be interpreted as a retrogression from a tritubercular tooth. In the same way may be explained the triconodont teeth of Seals and of the Cetacean Zeuglodon. Finally, the modern Toothed Whales have retrograded into "haplodonty."

Embryological evidence has also been called in, and with some success, to contribute towards the proof of the tritubercular theory of teeth. Taeker has shown that in the Horse and the Pig, and some other Ungulates, there is first of all a single hillock or cusp, and that later the additional cones arise separately. An apparent stumbling-block raised by these investigations is that it is not always the protocone or its equivalent in the upper jaw which arises first, as it obviously ought to do phylogenetically. This, however, is not a final argument in either direction. We know from plenty of examples that ontogenetic processes sometimes do not correspond in their order with phylogenetic changes. Thus in the mammalian heart the ventricle divides before the auricle; and of coarse, phylogenetically, the reverse ought to occur, since a divided auricle precedes a divided ventricle. This method of development has, moreover, been interpreted otherwise. It has been held to signify that the complex teeth of mammals are indeed derived from simple cones but by the fusion of a number of those cones.

On the other hand there are the claims of the multitubercular theory of the origin of mammalian teeth to be considered. The palaeontological evidence has been already, to some extent, utilised. The occurrence of such teeth among the possible forerunners of mammals, and in some of the most primitive types of Mammalia, has been referred to. Señor Ameghino dwells upon the sextubercular condition of many primitive mammals even belonging to the Eutheria. In a recent communication[^[30]] he attempts to identify six tubercles in the molars of types belonging to a

variety of Orders. The same condition, as has been noted, characterises that ancient Ungulate form Euprotogonia. Even where the teeth seem at first sight to be tritubercular a detailed study shows traces of otherwise vanished cusps.

It must be remembered in basing arguments upon the early Jurassic and Cretaceous mammals, that our knowledge of them mainly depends upon lower jaws, the teeth of which are usually simpler in pattern than those of the upper jaws. Moreover, another fact, not always insisted upon, must not be lost sight of. In many of those creatures the jaws were of small size, and yet accommodated a large series of molar teeth. Amphitherium, for example, had six molar teeth, and five is a number frequently met with. As the teeth are so numerous and the jaws so small it seems reasonable to connect the simplicity of the structure of the teeth with the need for crowding a number together. The same argument may partly account for the superabundant teeth of many Toothed Whales. It is true that the Manatee has very numerous grinders which are yet complex; but then in this animal there is a succession, and the jaw does not hold at a given time the entire series, with which it is provided in relays. On the other hand, where there are few molars they are often of the multitubercular type, or at least approach it; of this the Multituberculate Polymastodon is a good example; so, too, the molars of Hydrochoerus, and of many other Rodents.

It is well known that the fourth deciduous molar of the upper jaw, which is replaced by a permanent premolar in the fully adult animal, is of a more complex structure than its successor. This may indeed be extended to premolars earlier in the series. In the Dog "the second and first milk molars closely resemble the third and second premolars"; now the milk premolars belong evidently to the same dentition as the permanent molars, and they are earlier teeth than the later-developed replacing teeth. It is therefore significant that these earlier teeth should be more cuspidate than the later teeth. It tells distinctly in favour of the simplification as opposed to the complication of teeth in time, in the groups concerned.

These facts may possibly be applied in explanation of the simple teeth of some of the Jurassic and Cretaceous mammals. It has been mentioned that absolute trituberculy is exceedingly rare among those ancient creatures; more generally there are to

be found at least traces of more cusps. Now in some of them we may be dealing with instances of a complete tooth change; the suppression, save for one tooth, which is found in Marsupials, was probably not developed in at least some of these early mammals. The simplicity may therefore have been preceded by complexity, and may have been merely an adaptation to an insectivorous diet.

Alimentary Canal.—The mouth of the Mammalia is remarkable for the fact that with a few exceptions, such as the Whales, there are thick and fleshy lips. The office of these is to seize the food. The roof of the mouth is formed by the "hard palate" in front, which covers over the maxillary and palatine regions. This region is often covered with raised ridges, which have a symmetrical disposition, and are particularly strong in Ruminant animals. They are much reduced in the Rodents, where the anterior part of the palate is ill-defined owing to the way in which its sides fade into the lateral surface of the face. It has been shown that these ridges, in the Cat at least, develop as separate papilliform outgrowths, and it has been suggested that these papillae, which later become united to form the ridges, are the last remnant of palatine teeth such as occur in lower vertebrates.

Fig. 40.—Palatal folds of the Raccoon (Procyon lotor). p.p, Papilla palatina; r.p, palatal folds. (From Wiedersheim's Structure of Man.)

The tongue is a well-developed organ, usually playing a double part. It acts as an organ of prehension, especially in such animals as the Giraffe and the Anteater, where it is long and protrusible beyond the mouth for a considerable distance. It also carries gustatory organs, which serve for the discrimination of the nature of the food. Beneath the tongue there may be a hardish plate, known as the sublingua. This is especially prominent in the Lemurs, where it projects as a horny structure below the tongue, and has an independent and free tip. It is supported in some of these animals by a cartilaginous

structure. It is held by Gegenbaur that this organ is the equivalent of the reptilian tongue, and that in the skeletal vestiges which it contains are to be found the equivalents of the hyoid skeletal cartilages which support the tongue in lizards. In this case the tongue of mammals is a subsequently added structure.

The oesophagus leads from the mouth cavity to the stomach. The latter organ has commonly a distinctive shape in mammals. This is well shown in Man. The orifices of the oesophagus and intestine are somewhat approximated; and this causes a bulging of the lower border of the organ, usually spoken of as the greater curvature. A stomach of this typical form is found in many orders of mammals, and is unlike the stomach in any of the groups of lower vertebrates in shape. Sometimes the shape of the organ is greatly altered: it may be drawn out, sacculated, or divided, as in the Ruminants and Whales, into a series of differentiated chambers, each of which plays some special part in the phenomena of digestion.

The intestine of mammals is always long and much coiled, though the length and consequent degree of coiling naturally varies. On the whole it is perhaps safe to say that it is shorter in carnivorous than in vegetable-feeding beasts. Thus the Paca has an intestine of 39 inches total length, while the Cat, an animal of about the same size, has an intestine which is only 36 inches long. A fish diet, however, to judge from the Seals, is associated with a long intestinal tract. The intestine is divisible in the vast majority of mammals into a small and a large intestine. The two are separated by a valvular constriction save in certain Carnivores; and in the majority of cases the distinction is also emphasised by the presence at the junction of a blindly-ending diverticulum, the caecum. This latter organ varies greatly in length, being very short in the Cat-tribe and exceedingly long in Rodents. Its size is, to some extent, dependent upon the flesh-eating or grass-eating propensities of the animal in which it occurs. One of the longest caeca is possessed by the Vulpine Phalanger, in which the organ is one-fifth of the length of the small intestine; while the opposite extremity is reached by Felis macroscelis, which has a small intestine one hundred times the length of the caecum.

Fig. 41.—Different forms of the stomach in Mammals. A, Dog; B, Mus decumanus; C, Mus musculus; D, Weasel; E, scheme of the Ruminant stomach, the arrow with the dotted line showing the course taken by the food; F, Human stomach. a, Minor curvature; b, major curvature; c, cardiac end. G, Camel; H, Echidna aculeata. Cma, Major curvature; Cmi, minor curvature. I, Bradypus tridactylus. Du, Duodenum; MB, coecal diverticulum; **, outgrowths of duodenum; †, reticulum; ††, rumen. A (in E and G), Abomasum; Ca, cardiac division; O, psalterium; Oe, oesophagus; P, pylorus; R (to the right in E and to the left in G), rumen; R (to the left in E and to the right in G), reticulum; Sc, cardiac division; Sp, pyloric division; WZ, water-cells. (From Wiedersheim's Comparative Anatomy.)

An interesting point in connexion with the gut of mammals is the varying proportion of the small to the large intestine. As a general rule the former is very considerably longer than the latter; in Paradoxurus, for instance, the small intestine may be fifteen times the length of the large. The excess of length of one section over the other is not generally so marked

as this. In Phalanger maculatus the two sections of the gut are as nearly as possible equal in length, while in Phaseolarctos the large intestine is considerably longer than the small, the lengths being respectively 160 inches and 111 inches. It is common among the Marsupials and also among the Rodents for these proportions to exist, i.e. for the large intestine to be as long as, or longer than, the small. But there are so many exceptions that no general statements can be extracted from the facts.

Some few details will be found in the systematic part of this book. Mr. Chalmers Mitchell has brought forward some reasons for associating a great length of large intestine with an archaic systematic position, in the birds at any rate. The facts here briefly touched upon are not at variance with the extension of such a view to the mammals.

Fig. 42.—Diagrammatic plan of the liver of a Mammal (posterior surface). c, Caudate lobe; cf, cystic fissure; dv, ductus venosus; g, gall-bladder; lc, left central lobe; ll, left lateral lobe; llf, left lateral fissure; p, portal vein entering transverse fissure; rc, right central lobe; rl, right lateral lobe; rlf, right lateral fissure; s, Spigelian lobe; u, umbilical vein; vc, post-caval vein. (After Flower and Lydekker.)

Appended to the alimentary tract are three glands or sets of glands. Opening into the mouth cavity are the salivary glands, which are of enormous size in Anteaters, and small or absent in Whales. In their number and position these glands are characteristic of mammals. Into the intestine open the ducts of the pancreas and liver, two glands which the mammals share with lower vertebrates. The form of the liver is, however, generally characteristic of mammals. It is divided as a rule into a right and a left half, the line of division being marked by the insertion of the umbilical ligament, a vestige of the primitive ventral mesentery. Each half is again commonly subdivided into central and lateral lobes. In addition to these, two other divisions are often to be seen—the Spigelian and the caudate lobe. The liver is less divided in Cetacea and

some others, very much subdivided in Rodents and other groups. The degree of subdivision and the proportions of the several lobes frequently offer valuable systematic characters. The gall-bladder may be present or absent; it is always a diverticulum of the hepatic duct. The two are never separate, as in birds, for instance.

Organs of Circulation.—The heart of all mammals is a completely four-chambered organ. In the adult heart there is no communication between the right and left halves. The auricles are comparatively thin-walled, the ventricles thick-walled, in relation to the amount of work that they have severally to perform. The right ventricle, moreover, which has only to drive the blood into the lungs, is much thinner-walled than the left ventricle, which is concerned with the entire systemic circulation. The exits of the arteries and the auriculo-ventricular orifices are guarded by valves, which are so arranged as only to permit the blood to flow in the proper direction. But these valves have a morphological as well as a physiological interest. At the origin of each artery, the aorta and the pulmonary, there is a row of three watch-pocket valves, as they have been generally termed on account of their form. These three valves meet accurately in the middle of the lumen of the arterial tube when liquid is poured into them from above, and thus completely occlude the orifice. The auriculo-ventricular valves differ in structure in the two ventricles. That of the left ventricle has only two flaps, and is therefore often spoken of as the bicuspid or mitral valve. Both these flaps are membranous, and together they completely surround the exit from the auricle into the ventricle. The edges of the valve are bound down to the parietes of the heart by numerous branching tendinous threads, the chordae tendineae, which often take their origin from pillar-like muscles arising from the walls of the heart, the so-called musculi papillares. The valve of the right ventricle is composed of three flaps, and is therefore often spoken of as the tricuspid valve; it is in the same way membranous, and has chordae tendineae and musculi papillares connected with it. The disposition of the musculi papillares and their number differ in different mammals, but no exhaustive study has as yet been made of the arrangements in different groups; the amount of individual variation even is not known, though it is certainly considerable in some cases, for

instance in the heart of the Rabbit. The heart of the Monotremata presents differences of some importance from those of other Mammalia; the modern knowledge of the Monotrematous heart is mainly due to Gegenbaur[^[31]] and Lankester,[^[32]] in whose memoirs references to the older literature will be found. The principal features of interest in which the heart of the Monotremata differs from that of the higher Mammalia are these. When the two ventricles are cut across transversely, the cavity of the right is seen to be wrapped round that of the left in a fashion precisely like that of the bird's heart; on the other hand in the higher mammal the two cavities lie side by side. The main difference between Monotremes and other Mammals concerns the right auriculo-ventricular valve. The differences which it presents from the corresponding structure of the rest of the Mammalia are two: in the first place, the valve itself does not completely surround the ostium; it is only developed on one side; the septal half (i.e. that turned towards the interventricular septum) is either entirely absent or more generally represented by a small bit of membrane; nevertheless I found[^[33]] recently in an Ornithorhynchus heart a complete septal half to the right auriculo-ventricular valve. The second point of interest in connexion with this valve is, that the musculi papillares instead of ending in chordae tendineae attached to the free edge of the valve are directly attached to the valve, and in some cases pass through its membranous flap, to be attached to its origin at the boundary of the auricle and of the ventricle. The invading of the valve-flap by muscle in this way is highly interesting, as it recalls the heart of the bird and of the crocodile. The imperfect condition of the valve (from which, as has already been stated, the septal half is as a rule nearly absent) is a point of resemblance to the heart of the bird; the corresponding valve of the crocodile's heart being complete.

Fig. 43.—Lepus cuniculus. Ventral view of the vascular system. The heart is somewhat displaced towards the left of the subject; the arteries of the right and the veins of the left side are in great measure removed. a.epg, internal mammary artery; a.f, anterior facial vein; a.m, anterior mesenteric artery; a.ph, anterior phrenic vein; az.v, azygos vein; br, brachial artery; c.il.a, common iliac artery; c.il.v, common iliac vein; cœ, coeliac artery; d.ao, dorsal aorta: e.c, external carotid artery; e.il.a, external iliac artery; e.il.v, external iliac vein; e.ju, external jugular vein; fm.a, femoral artery; fm.v, femoral vein; h.v, hepatic veins; i.c, internal carotid artery; i.cs, intercostal vessels; i.il.a, internal iliac artery; i.il.v, internal iliac vein; i.ju, internal jugular vein; i.l, iliolumbar artery and vein; in, innominate artery; l.au, left auricle; l.c.c; left common carotid artery; l.pr.c, left pre-caval vein; l.v, left ventricle; m.sc, median sacral artery; p.a, pulmonary artery; p.epg, epigastric artery and vein; p.f, posterior facial vein; p.m, posterior mesenteric artery; p.ph, posterior phrenic veins; pt.c, post-caval vein; p.v, pulmonary vein; r, renal artery and vein; r.au, right auricle; r.c.c, right common carotid artery; r.prc, right pre-caval vein; r.v, right ventricle; s.cl.a, right subclavian artery; s.cl.v, subclavian vein; spm, spermatic artery; s.vs, vesical artery; ut, uterine artery and vein; vr, vertebral artery. (From Parker's Zootomy.)

There are also features in the system of arteries and veins which are eminently distinctive of mammals. In the first place, the aorta leaving the heart and conveying blood to the body is only a half arch, and bends to the left side as seen in Fig. 43. The right and left halves are present in reptiles, and meet behind the heart. In the bird the right half alone has remained. This fact, therefore, shows that the mammal cannot have been derived from a bird-like ancestor, but that both must have independently come from an ancestor with both halves of the aortic arch present, of which one half has disappeared in one group, and the other half in the other. It is an interesting fact, too, to notice that the four

cavities of the mammal's heart, which fourfold division it shares with birds alone, do not exactly correspond compartment for compartment with those of the bird's heart, at least in so far as concerns the ventricles. For the reptilian heart is provided with only one ventricle, and therefore the division of that cavity must have been independently accomplished in mammals and in birds.

There are two features in the venous system which distinguish all the Mammalia (with the exception of Echidna in one of these points) from vertebrates standing lower in the series. The hepatic portal system is limited to a vein which conveys to the liver blood derived from the alimentary tract; in no mammal except in Echidna is there any representative of the anterior abdominal vein of lower vertebrates. In that animal there is such a vein, which apparently arises from a capillary network upon the bladder and passes up, supported by a membrane, along the ventral wall of the abdomen to the liver, thus emptying blood into that organ exactly as does the anterior abdominal vein of the frog. In no mammal is there any trace of a renal portal system. The kidneys derive their blood from the renal arteries only.

Many mammals have two superior venae cavae; this is the case, for instance, in the Elephant and the Rodents and other types lying comparatively far down in the series. In most if not in all mammals there are considerable remains of one of the posterior cardinals, in the form of the azygos vein, which opens into the vena cava superior or pre-caval vein, i.e. the superior cardinal just before the latter debouches into the heart. This one posterior cardinal is usually on the right side; but it may be on the left side, for instance in Trichosurus vulpecula. In Halmaturus bennettii there are two azygos veins, one left and one right, of which the left is rather the larger.[^[34]]

Urinary Organs.—The kidneys in the Mammalia have a compact form, which contrasts with the somewhat diffuse and vaguely-outlined kidneys of the Sauropsida. In mammals the organ is as a rule of that peculiar shape which is called "kidney-shaped"; a depression termed the hilum, which receives the ducts of the glands, indenting the border of an otherwise oval-shaped gland. In some few mammals the kidney is broken up

into lobules; this is the case with the Whales, the Bears, the Oxen, and a few other forms. A curious fact about the kidneys of the Mammalia is their very general asymmetry of position. One of them usually lies in a more advanced position than the other. The ureters lead from the kidneys to the urinary bladder, which in its form and relations is quite distinctive of the Mammalia. The bladder is formed out of the remains of the allantois, and is therefore not the exact homologue of the bladder of the frog, which is the equivalent of the entire sac which grows out of the cloaca in the mammal, and is the foetal allantois. The ureters open into the bladder in the higher Mammalia, but lower down in the urino-genital passage in the more primitive mammals.

The Body Cavity.—The Mammalia differ from all other living vertebrates by the arrangement of the body cavity in which lie the viscera. That cavity is divided into two by a partly muscular and partly tendinous partition, the diaphragm. No other vertebrate has this precise disposition of the coelom. The diaphragm lies usually transversely to the longitudinal axis of the body, but gets a much more oblique arrangement in the Cetacea and the Sirenia, whose needs demand a more expanded chamber for the lungs. For in front of the diaphragm lie the lungs and heart; behind it the stomach, liver, intestines, and the organs of reproduction and excretion. The diaphragm is used in respiration; when its muscles contract, the surface directed toward the pleural cavity becomes less convex, and the cavity of the lungs is thus increased, allowing them to expand under the pressure of the entering air.

The Lungs.—The lungs of the Mammalia differ from those of animals lying lower in the series by the fact, just referred to, that they occupy a pleural cavity completely shut off from the abdomen by the diaphragm. As a rule the lungs of the Mammalia are to be distinguished by their more or less extensive lobation. In the Whales, however, and in the Sirenia, they are not much divided, but present the appearance of the simple sac-like lungs of the reptiles. In some mammals there is a median and posterior unpaired lobe of the lung, which lies in the post-pericardial cavity behind the pericardium. This is not universally present. The lungs are very frequently not symmetrical in their lobation, the number of separate lobes on the right side

and on the left being different. The lungs of mammals agree with those of the lower reptiles in being freely suspended within their coelomic cavity, and in not being, as in birds, crocodiles, and the Varanidae among lizards, tied down to the dorsal surface of that cavity by a sheet of peritoneum covering them.

Fig. 44.—Part of a sagittal section of an ovary of a child just born. bl.v, Blood-vessels; foll, strings and groups of cells derived from the germinal epithelium becoming developed into follicles; g.ep, germinal epithelium; in, ingrowing cord of cells from the germinal epithelium; pr.ov, primitive ova. (From Hertwig, after Waldeyer.)

The Gonads (Ovaries and Testes).—The ovary in the Mammalia is always paired; there is never a partial or complete abortion of one gonad as in birds—except of course in pathological cases. The ovaries are small, and lie in the abdominal cavity behind the kidneys. In the immense majority of the Mammalia the ova which are produced within the ovaries are of minute size; those of even the colossal Rorqual are, so far as we know, not markedly larger than the ova of a Mouse. The smallness of size of these reproductive elements implies necessarily an absence of much nutritive yolk; and as a consequence the developing embryo, since it is not hatched in an early stage as a free living larva, has to be nourished by the mother, to whose tissues it is attached through the intermediary of the placenta, a structure partly composed of foetal structures derived from the embryo, and partly of portions of the lining membranes of the uterus of the mother. The ova of the

Eutherian mammals, including the Marsupials, are very small as compared with those of any other vertebrates, excepting only Amphioxus, where the young are hatched early as free swimming larvae. They also differ in a highly characteristic way in the mode of their development within the ovary. These processes are to some extent illustrated in Fig. 44. The main framework of the ovary is formed of the so-called "stroma," which is a mass of tissue formed of more or less connective-tissue-like cells. Within this are numerous cavities, the Graafian follicles. The very young follicles consist of but a single layer of follicular cells surrounding the ovum, which lies centrally. The follicular cells gradually increase in number until the ovum lies in the midst of several layers of cells. At this period a vacuity is formed between some of these cells, and grows into a large cell-free cavity; the ovum does not lie loosely in this space, but is connected at one side with the follicular cells, which still line the interior of the Graafian follicle by the so-called discus or cumulus proligerus. The egg or ovum has, moreover, a layer of cells immediately surrounding itself. All these facts can be gathered by an inspection of Fig. 45. It has been shown that, as in lower vertebrates, the cells immediately surrounding the ovum are connected with it directly by delicate processes which penetrate the actual membrane of the egg.

Fig. 45.—Two stages in the development of the Graafian follicle. A, With the follicular fluid beginning to appear; B, after the space has largely increased. caps, Capsule; disc, cumulus proligerus; memb, membrana granulosa; ov, ovum; sp, space containing fluid. (After Hertwig.)

Fig. 46.—Ovarian egg of Echidna. b, Basilar membrane; fe, follicular epithelium; o, oil globules; vm, vitelline membrane; y¹, y², yolk. (Partly after Caldwell.)

The only ova which depart at all in structure from that above described are those of the Monotremata. The credit of this

discovery rests with Owen and with Professor Poulton, who pointed out in 1884,[^[35]] that the ovum of Ornithorhynchus is very large as compared with those of other Mammalia (6 mm. as against .2 mm.), that it is filled with yolk, and that it completely fills the follicle, being surrounded by two layers of follicular cells only. This latter fact was proved by Caldwell. Subsequently Gyldberg[^[36]] and I[^[37]] described the ovarian ovum of Echidna, showing it to be identical with that of Ornithorhynchus. Later still a more elaborate and beautifully illustrated paper was published by Caldwell[^[38]] upon the early stages of development in the Monotremata and Marsupials, in which the ovum of the former was accurately described (see Fig. 46). In the particulars mentioned above, the ovum of the Monotremata is practically identical with that of the large-yolked ova of the Sauropsida.

Fig. 47.—Lepus cuniculus. The anterior end of the vagina, with the right uterus, Fallopian tube, and ovary. (Nat. size.) Part of the ventral wall of the vagina is removed, and the proximal end of the left uterus is shown in longitudinal section, fl.t, Fallopian tube; fl.t′, its peritoneal aperture; l.ut, left uterus; l.ut′, left os uteri; ov, ovary; r.ut, right uterus; r.ut′, right os uteri; s, vaginal septum; va, vagina. (From Parker's Zootomy.)

It is the general rule among vertebrate animals that the ovaries are completely independent of the ducts which convey their products to the exterior. In certain fishes, however, there is an absolute continuity between the two structures, which is believed to be due to a simple concrescence between the originally distinct ovary and oviduct. The latter has grown round the former, an obvious advantage in preventing the eggs from wandering into the abdominal cavity and becoming lost. In the Mammalia we find discontinuity as a general rule. But in quite a number of forms folds of the lining membrane of the abdominal cavity are developed, which practically ensure the passage of the ova into the oviduct when they are extruded from the ovaries. The oviduct, moreover, has a large and fimbriated mouth, called in human anatomy—which is provided with a number of fanciful names—the morsus diaboli. This almost wraps round the ovary, and thus prevents the ova from straying in the wrong direction. Moreover, the ovary itself is often so arranged that it can easily be withdrawn into a pocket of the peritoneum, from which the obvious exit is by the gaping mouth of the oviduct. This disposition of the generative parts is still further modified in a few animals, such as the Rat[^[39]] and the Kinkajou.[^[40]] In these animals the mouth of the oviduct actually opens into the interior of a closed chamber which contains the ovary. In this case there is but one route for the

extruded ova to follow. This series of steps in the perfecting of the mode of safe extrusion of the ova is highly interesting, and is a piece of evidence in favour of the high position of the mammals.

Fig. 48.—Female urino-genital apparatus of various Marsupials. A, Didelphys dorsigera (young); B, Trichosurus; C, Phascolomys wombat. B, Urinary bladder; Cl, "cloaca"; Fim, fimbriae; g, clitoris; N, kidney; Od, Fallopian tube; Ot, aperture of Fallopian tube; Ov, ovary; r, rectum; Sp, septum dividing vagina; Sug, urino-genital sinus; Ur, ureter; Ut, uterus; Ut′, opening of the uterus into the median vagina (VgB); Vg, lateral vagina; Vg′, its opening into the urino-genital sinus; † (in B), point of approximation of uteri; † (in C) and *, rectal glands. (From Wiedersheim's Comparative Anatomy.)

The oviducal apparatus of the mammal is more specialised than that of lower vertebrates. It is most simple, as might be imagined, in the egg-laying Monotremes, where, indeed, it is on the same level as that of reptiles. But in the Eutheria the fimbriated mouth of the oviduct passes into a narrow and winding tube, the Fallopian tube; this widens into a uterus, and the two uteri combine into a single tube in the higher forms. They are called the Monodelphia on this account. In the Marsupials the uteri are distinct though they often join above, and from this junction depends a median "uterus." After the uterus or the uteri follows in every case a single vagina.

The testes of the Mammalia, like those of other vertebrates, occupy primitively a position within the body cavity precisely corresponding to that of the ovaries. And in the lowly-organised Monotremata, and some other forms, such as the Whales, they retain that primitive position within the body. It is, however, distinctive of the Mammalia as opposed to lower vertebrates that the testes descend later into a scrotum, which is simply a protrusion of the skin of the body surrounded by muscles, and, of course, containing a section of the body cavity in which lie the testes. The penis of the Mammalia, represented by the clitoris and associated structures in the female, is of a structure entirely peculiar to this group.

Fig. 49.—Brain of Dog. A, ventral; B, dorsal; C, lateral aspect. B.ol, Olfactory lobe; Cr.ce, crura cerebri; Fi.p, great longitudinal fissure; HH, HH¹, lateral lobes of cerebellum; Hyp, hypophysis; Med, spinal cord; NH, medulla oblongata; Po, pons Varolii; VH, cerebral hemispheres; Wu, middle lobe (vermis) of cerebellum; I-XII, cerebral nerves. (From Wiedersheim's Comparative Anatomy.)

The Brain.—Inasmuch as Professor Wiedersheim has said with perfect truth that "the brain of the extinct Ungulate Dinoceras shows so striking a likeness to that of a lizard that one would be compelled to explain it as that of a lizard without a knowledge of the skeleton," it is clear that to define the mammalian brain is a difficult matter. The existing Mammalia, however, all possess brains which can be readily distinguished from those of vertebrates lying lower in the scale. They are of relatively large size, brought about mainly by the dimensions of the cerebral hemispheres, which have an importance in this class of vertebrates that they have not elsewhere. Coupled with this large size of the hemispheres is a more elaborate system of transverse commissures uniting the two; and this culminates in the higher Mammalia, where the corpus callosum attains a large size and great physiological importance. A

very marked feature, moreover, of the mammal's brain is the development of regular fissures upon its surface, which fissures are only absent from Ornithorhynchus, various small Rodents, Bats, and Insectivores, among living mammals. It is sometimes, but erroneously, said that the more complicated the fissures of the brain are, the higher in intelligence and "zoological position" is the possessor of that brain. Instances can undoubtedly be quoted to support such a view; but they are

merely selected cases, which do not indicate a wide applicability of such a generalisation. Thus it is true that the brain of a Man is more elaborate in its furrows and convolutions than is that of a Cat. The real fact of the matter is, that the complexity of the brain from this point of view increases with the size of the animal within the group.

Fig. 50.—Lepus cuniculus. Longitudinal vertical section of the brain. (Nat. size.) a.co, Anterior commissure; b.fo, body of the fornix; cb, cerebellum, showing arbor vitae; c.c, crus cerebri; c.h¹, parencephalon or cerebral hemisphere; c.h², temporal lobe; c.ma, corpus mammillare; cp.cl, corpus callosum; f.m, foramen of Monro; inf, infundibulum; l.t, lamina terminalis; ly, lyra; m.co, middle commissure; m.o, medulla oblongata; o.ch, optic chiasma; o.l¹, o.l², corpora quadrigemina or optic lobes; olf, olfactory lobe; p.co, posterior commissure; pd.pn, peduncle of the pineal "gland," pn; p.fo, anterior pillar of the fornix; pty, pituitary body; pv.a, pons Varolii; sp.lu, septum lucidum; v⁴, fourth ventricle; vl.ip, velum interpositum; v.vn, valve of Vicussens; II, optic nerve. (From Parker's Zootomy.)

The Gorilla and the Chimpanzee have a more furrowed brain than has the little Marmoset; the Bear a more complicated brain than the Weasel, etc. The most highly-convoluted brains of all mammals are those of the Elephants, and there does not seem in the Ungulates to be so marked a relation between size and abundance of fissures as there is among other mammals. A regular plan of the fissures can be detected with certainty for each group considered by itself; but it is not so easy to homologise the details of arrangement from group to group. This is so far in accord with the view that the existing groups of mammals have diverged from each other ab initio.

Another marked characteristic of the mammalian as opposed to other brains is the relatively small importance in size and yet the fourfold nature of the optic lobes. What was the case with the optic lobes of the early Ungulates is difficult to understand, on account of the fact that the casts are necessarily imperfect.

Altogether the enormous progress in the complexity of the brain from the early Tertiary mammals down to the present, is one of the most remarkable revelations of palaeontology. It goes perhaps some way in explaining the remarkable diversity in mode of life exhibited by the mammals as compared, for example, with the birds, whose brains have not diverged so much or in so many directions from the primitive form.

The present Distribution of the Mammalia.—In the following pages some of the principal facts in the geographical range of the orders, families, and many of the genera of Mammalia will be given. It has been justly observed by Mr. Sclater that the habitat of an animal is as much a part of its definition as is its structure or external form. No systematic account of the Mammalia would therefore be complete without such geographical facts. But that branch of zoology which is concerned with the past and present distribution of animals is wider in scope than this. Zoogeography deals not only with the actual facts in the range of animals, but with the inferences as to past changes in the relations of land and sea which the facts seem to indicate, and with speculations as to the place of origin of the different groups, of which more than hints are sometimes given by their past and present distribution. In addition to this, the earth can be mapped out into provinces and regions which are definable by their animal inhabitants. In the present volume, dealing only with the Mammalia, it will be obviously impossible to enter fully into the entire subject of zoogeography. All that will be attempted is a brief general survey of the science so far as it can be illustrated by the Mammalia. For fuller knowledge the reader is referred to the treatises mentioned below.[^[41]]

There are certain facts in the distribution of animals which are commonplaces of knowledge, but which may be set forth with definiteness. Everybody knows that an animal has a given range: Elephants, for example, are found in India and certain adjacent parts of Asia, and again in Africa; the Rhinoceroses have roughly the same range; the Tiger is limited to Asia; the

Jaguar to America, and so forth. The entire expanse of country which is inhabited by an animal is called its area of distribution. Such areas are larger or smaller. The Lion ranges over the whole of Africa, a small part of India, and some neighbouring countries; on the other hand, the Insectivore Solenodon is limited to Cuba and Hayti, a separate species to each. Among other groups of animals are instances of an even more restricted range. There are humming-birds confined to the slopes of a single mountain, and fishes limited in their range to a single small lake.

A species may be found everywhere within the area of its distribution, or it may be confined to a number of limited tracts within that area. In this case it is usual to speak of "stations." In such cases the species in question is generally suited to some particular kind of environment. Thus the Otter and other aquatic mammals will only be found where there is water; and intervening tracts of waterless country will contain no Otters. Goats and Chamois live only upon mountains; the intervening plains are destitute of them. This discontinuity of distribution within the area is very general. But a discontinuity of area is also seen—not so commonly however; and, indeed, when it does occur, it is a matter of a genus and not of a species. Thus the Tapir is found in the East Indies on the one hand and in South and Central America on the other, being absent in the intermediate tracts.

It is clear that tracts of country eminently suitable for the housing of a particular mammal do not always possess that kind, or even an allied form. Africa, for example, possesses no arboreal Anteaters; there are no Anteaters at all (of the order Edentata) in Australia, though there are plenty of ants for them to feed upon, and tropical conditions of climate prevail. But as in these cases the inference may be denied on the grounds that no experiments exist to prove or to disprove the assertion, the matter may be better emphasised by such cases as the introduction of the Rabbit into Australia, and various mammals, such as Goats, into oceanic islands. The plague caused by the former is a matter of notoriety. But although climate and conditions and animal inhabitants do not march accurately together, there is certainly some connexion between temperature and the range of animals. Mr. Lydekker writes on this point as follows: "The llama-like animals, respectively known as vicunas and guanacos, are met with in

company on the highlands of the Cordillera in Peru and Ecuador, but as we go farther south the latter are found on the plains of southern Argentina and Patagonia, as well as on the island of Tierra del Fuego at the sea level. Here then is a clear proof of the intimate connexion existing between temperature and station; the guanaco being an animal which can only live in cold or temperate climates, finds suitable conditions for its existence in tropical latitudes solely at a height of so many thousands of feet, although farther south it is able to thrive at the sea level." This, however, cannot be pushed too far—the world cannot be mapped out into areas bounded by parallels of temperature as was once attempted—since there are plenty of cases like that of the Tiger, which is as much at home in a tropical jungle as on the icy plains of Northern Asia.

Seeing that there are in many cases no climatic barriers to the spreading of a given race of animals over a larger area of distribution than it actually occupies, it becomes important to inquire why there are so many cases of restriction in range.

It is possible to see, at any rate, three causes which are responsible for a large number of such cases. In the first place, a given species of animal must have originated at a certain spot; its multiplication in individuals must always be a slow matter, since enemies, and untoward events generally, would conspire to check the natural multiplication by geometrical progression. A long time might therefore elapse before the species greatly extended its range. A restricted distribution may therefore, in some cases, mean a modern race. In the second place, there are definite physical barriers which check the migration of species. The terrestrial Mammalia cannot cross wide arms of the sea; that they can and do swim for considerable distances has been proved in several instances; but, as has been pointed out, it is unlikely that a purely terrestrial mammal would voluntarily swim out into an unknown sea. And then if it did, and successfully reached the opposite side, nothing would happen unless it were a pregnant female; or, if not pregnant, till a male swam very soon afterwards in exactly the same direction. Many travellers have told of floating islands, formed of torn-up trees and brushwood, which have been seen at the mouths of large rivers, with animal passengers upon them. These are, however, so much at the mercy of currents and storms,

that but little reliance can be placed on them as a means of transit; besides, here again, two individuals, or a pregnant female, would be required to effect a settlement on a foreign shore. The existence of oceanic islands is often urged as a proof of this inability to cross tracts of sea; even those which are comparatively near an extensive continent, such as, for example, Fernando Noronha in the Atlantic, are destitute of mammals (except, indeed, the ubiquitous Mouse, which is believed to have been carried there, often in company with the equally widely-spread Rat, in ships). This argument, however, is not so conclusive as might appear; it doubtless is in the case of far-distant islands. But the size of the islands has to be taken into account. For there are islands, such as the Galapagos, or, to take a less contested instance, some of the islands of the Malagasy Archipelago, undoubtedly continental, which have an exceedingly reduced number of mammals. An area of a certain size seems to be a necessity.

The converse of this is in many cases easy to show, that is, the wide range of animals when there are no marine barriers to stop their spreading. John Hunter, the celebrated anatomist and surgeon (not often quoted, however, as an authority upon geographical distribution), observes: "It is a curious circumstance in the natural history of animals to find most of the northern animals the same both on the continent of America and what is called the Old World, while those of the warmer parts of both continents are not so. Thus we find the bear, fox, wolf, elk, reindeer, ptarmigan, etc., in the northern parts of both.... The reason why the same animals are to be found in the northern parts is the nearness of the two continents. They are so near as to be within the power of accident to bring the animals, especially the large ones, from one continent to the other either on the ice or even by water. But the continents diverging from each other southward, so as to be at a very considerable distance from each other even beyond the flight of birds, is the reason why the quadrupeds are not the same."

There is no doubt, in fact, that the ocean is the most insuperable of all barriers to the dispersal of mammals. In a less degree mountain ranges and deserts are also barriers. The Desert of Sahara is a striking instance to the point; it separates two exceedingly different faunas.

A third cause of more or less limited range is the barrier due to competition. If the ground is already taken up, there is no room for new immigrants. There is obviously a limit to the number of Antelopes or Deer that can graze upon a given tract of grassy plain. These two groups of Ungulates illustrate the matter well: the Antelopes are African and Indian, especially the former, while Africa has no Deer at all; America, on the other hand, has plenty of Deer but no Antelopes, save the Prong-horn. The more nearly akin the two species or groups of species are, the fiercer will be the competition; for a near kinship will at least often imply similar habits, the need for similar food, and other likenesses which will prevent both from successfully occupying the same tract of country. The remarkable fauna of Australia is believed to afford an example of this. In that country the prevalent inhabitants are the Marsupials. The Monotremes are found there also, and nowhere else save in New Guinea and Tasmania. The remaining mammals are inconspicuous; they embrace a few Rodents and Bats, and the doubtfully indigenous Dingo-dog. Now the Marsupials are fitted to every variety of life. We have the grazing Kangaroos and Wallabies, the burrowing Wombats, the arboreal Phalangers, and the carnivorous Dasyures. In the second place, it is an unquestioned fact that the Marsupials are an older race than are the existing Eutherian mammals; they were the dominant mammals during the Secondary epoch. At that time they were more widely distributed than at present. In most parts of the world they are now absent, since they have been successfully ousted by the more highly organised groups of Eutheria. But at that period, when the higher Eutheria were in the ascendant, Australia and the islands to the north became cut off from Asia, and thus became freed from inroads of Eutheria, which were partly prevented by the physical barrier of the sea from effecting a settlement, and partly perhaps prevented owing to the ground being already taken up by the Marsupials. Likeness of habit gave the older inhabitants victory in the struggle for existence.

The general statements that have been here made are in accord with current opinion upon the factors of geographical distribution. But the past range of animals appears to be less consonant with the received views. In the Tertiary

period, groups of animals had often a far wider range than at present. To-day the Rhinoceroses are limited to Asia and Africa, and to quite limited parts of the former continent. In the past, these animals were abundant in Europe and North America. Wild Horses now have a range which is not widely different from that of the Rhinoceroses, save that they extend into the more northern regions of Asia. Their remains are abundant both in North and South America. The Hippopotamus, now confined to Africa, once ranged over Europe, Madagascar, and India. There were plenty of American and European Lemurs. Elephants were nearly world-wide in their range; and, in short, restricted distribution seems to be on the whole a characteristic of animals of the present day.

These statements, however, though perfectly true, must not lead to erroneous inferences. It is rather impressed upon the reader, in books which contain sections dealing with geographical distribution, that animals on the whole occupy more restricted areas at present than in the past. There are, however, plenty of examples of groups of extinct creatures which had, so far as we know, quite a restricted range. Thus the Toxodonts were purely South American, as were the Glyptodonts and some other forms. And, on the other hand, the Cervidae of to-day are as widely, if not more widely, distributed than at any other time. The Hares and Rabbits are now nearly universal in range; the Cats almost so. We meet with Bovidae, even excluding the Sheep and Goats, in all the four quarters of the globe, excluding only South America and, of course, Australia. The Camelidae are still common to both the Old and the New Worlds.

During certain periods of the Tertiary epoch it is true that there was more similarity between Europe and North America than there is at present. It would have been quite necessary to unite both into a Holarctic area, such as is now insisted upon by many; but the reasons for this union would then have been stronger. The fact is, however, that the closer resemblances were due to the larger number of families of animals which existed then than now; these have decayed away from both continents, and allowed the unlikenesses between the mammalian fauna of both to become evident. But the likenesses which still survive have led many to associate the two regions closely together.

So far as the history of a genus or family or larger division

can be traced, it results as a conclusion that from a given area of origin the group in question migrated in all directions where possible to a varying degree; it then died out in intervening tracts, or was left only in a certain part of its former and more extensive area of range.

Zoological Regions.—Seeing that each species of animal has its own definite range, it is clear that the earth's surface can be apportioned into divisions which are characterised by their animal inhabitants. We shall divide the earth into realms, which are the largest divisions; then into regions; and finally into subregions. It must be borne in mind that the various groups of the animal kingdom are of different ages, geologically speaking, and have therefore had less or more time, as the case may be, to settle down into their present distribution, and that different animals differ greatly in their rate of multiplication, their power of migration, and their susceptibility to the effectiveness of various natural and other barriers to distribution. It is not, therefore, possible to divide the world into realms and regions which shall express the facts of distribution of the entire animal kingdom. Such divisions, which are common in text-books of zoology having but a small section devoted to zoogeography, are at best mere approximations and averages; no good is gained by taking such a comprehensive view of the matter, as the essential object of subdividing the earth's surface is thereby lost sight of. The zoogeographical division of the earth which will be adopted here is that originally recommended by Dr. Blanford, and now accepted by a number of authorities. There are three "realms," to which a fourth may perhaps be added—though on negative grounds, and merely for the purpose of emphasising the parts of the world to which mammals have not gained access. The realms are again divisible into regions, at least in the case of one of them, and the regions may be again separated into more or less distinct subregions or provinces. The three primary divisions or realms which contain mammals are the Notogaean, including Australia and certain islands to the north of it; the Neogaean, or the South American continent and Central America; the Arctogaean, including the continents of North America, Europe, Asia, and Africa, together with the adjacent islands, such as the West Indies, East Indies (exclusive of those which fall within

the realm of Notogaea), and Madagascar; and finally, the realm of Antarctogaea or Atheriogaea, which embraces New Zealand, the Antarctic continent, and a series of islands such as South Georgia and Kerguelen, and possibly even the extreme south of Patagonia. This latter quarter of the globe will need no further reference, as it has no truly indigenous terrestrial mammalian inhabitants. We cannot include the Bats in this statement, as their distribution is due to different powers of extending their range, and to different barriers from those which govern the range of other groups of mammals.

(1) Notogaea.[^[42]] This realm is characterised by the exclusive possession of the Monotremes:—that is to say, one of the two primary divisions of the Mammalia is absolutely restricted to this area. It contains, moreover, the vast majority of the Marsupials. Further, the realm of Notogaea is to be distinguished by the entire absence of the higher mammals, with the exception of a few small Rodents. (The Bats are ignored for the reasons stated, and the Dingo is believed to have been an importation.) It cannot be disputed that this is a very distinctly-marked area of the earth's surface.

(2) Neogaea. The continent of South America has no Monotremes and only a few Marsupials, all of which, with the exception of Caenolestes, belong to the Polyprotodont division of that order, and to a peculiar family, Didelphyidae. The recent discovery of other fossil Marsupials, however, to some extent favours Huxley's view that Neogaea and Notogaea form one realm as opposed to the rest of the world. Besides this, Neogaea possesses the Edentata, which are found nowhere else;—that is, the division of the Edentata to which the name is now restricted by some authorities. It is also characterised by the nearly entire absence of the important order of Insectivora; and, as minor marks of distinction, by the absence of Antelopes, Oxen and Sheep, of the Ichneumon tribe, of Horses, and of Lemurs. It has the exclusive possession of the Hapalidae and Cebidae, and of several families of Rodents.

(3) Arctogaea. This vast realm is clearly capable of subdivision into four regions, which will be considered in detail later. In the meantime the points of likeness between these subdivisions is more marked than are either the resemblances or the

differences of any one of them to either of the two realms which have just been defined. The two realms that have been discussed retain their distinctness from each other and from Arctogaea for a considerable way back into the Tertiary period. It is not until we reach very early Tertiary times that Edentates are met with in North America; and then it cannot be regarded as absolutely settled that the Ganodonta are really the forerunners of the Armadillos, Sloths, etc. Nor do we find Marsupials in Europe until far back in time, and at a corresponding period in North America. Indeed the fauna of South America in late Tertiary times was even more distinct than it is now; for then we had confined to that region the Toxodonts, Glyptodonts, Macrauchenia, and other forms, while in Australia there were still Marsupials. In late Tertiary times Europe and India were by no means so distinct from Africa as they are to-day. North America does not resemble the Old World quite so much as the subdivisions of the Old World resemble each other; but, as will be pointed out later, there are and were very substantial agreements. The Elephants, Rhinoceroses, Giraffe, Hippopotamus, Orycteropus, are now distinctively African or Indian animals; but all these genera, or at least families (in the case of the Giraffe), have occurred in Europe during quite recent times. Lycaon indeed, now confined to Africa, is thought to have had a European origin from its occurrence in caves there. The Hyaena and the Lion, certain members of the Horse tribe, Apes, and other animals, were also but are not now European.

India again, and the Oriental region generally, once possessed the Hippopotamus, the Chimpanzee, Giraffidae, the Antelopes, Cobus, Hippotragus, Strepsiceros, and Orias, which are now purely African animals. It shares at present with the Ethiopian region the Catarhines, including the Anthropoid Apes, the Lemurs, Tragulina (the genus Dorcatherium is also known from fossils in India), Manis, Hyaena, the Cheetah, Elephant, Rhinoceros, and the Ratel. There is, in fact, no order of mammals which is now absent from one of these three regions though present in the others, save the Lemurs, and they occurred in past times in Europe. The Tapir of India is known fossil in Europe, and the latter continent had its Monkeys and even Anthropoids. On the other hand, North America is more distinct. It has no Lemurs, Apes, Elephants, Rhinoceroses, Tapirs, Old World Edentates

(Effodientia), Viverridae, Horses, or Antelopes, excepting Antilocapra, a type of a separate division of Bovidae. But since several of these groups have been represented in recent times, no primary line of division can be profitably drawn.

Arctogaea as a whole may be characterised by both negative and positive characters. As negative features may be mentioned;—the entire absence of Edentates (Necrodasypus of Filhol is rather doubtful, see p. [164], n.), though a few crept up into the Nearctic region from Neogaea during past times; and of Hapalidae, Cebidae, and Marsupials, except an Opossum in North America. This realm has, on the other hand, all the Lemurs, all the Insectivores with the exception of the West Indian Solenodon, all the Proboscidea, Rhinoceroses, Horses, Deer, Antelopes, the last group including the Oxen and a variety of other important families. It is in fact the headquarters of all the Eutheria with the exception of the Edentata and Marsupials.

The subdivisions of this realm have been variously effected. The classical subdivisions are of course those of Mr. Sclater, who would recognise (1) the Nearctic, North America; (2) the Palaearctic, including Europe, Northern Asia, and Japan; (3) the Oriental, including Asia south of the Himalayas and the islands of the Malay Archipelago as far east as the Australian region; and (4) the Ethiopian, i.e. tropical Africa and Madagascar. Some would alter this by uniting America and the north of the Old World into a Holarctic region, separating off the southern parts of the North American continent into a Sonoran region. To some, the claims of Madagascar to form a separate region are convincing. To distinguish the boundaries of the several regions is a difficult task; they dovetail into each other on the frontiers with the complex curves of a puzzle-map. The difficulty has been grappled with by the suggestion of intermediate transitional areas; but this proceeding really doubles the difficulty, for there are then two frontiers to delimit in each case instead of only one. The animal inhabitants must be expected to mingle somewhat at the lines of junction of one region with another.

The Sonoran region does not appear to us to have great claims to recognition. It shows a mingling of southern with northern forms exactly as might be expected. An Armadillo and Didelphys have, as it is believed, invaded it from the Neogaeic realm; it possesses also the South American genera, Dicotyles, Nasua,

Conepatus, Sigmodon. On the other hand, the Sonoran genera Antilocapra, Cynomys, Procyon, and the Insectivora Blarina and Scapanus, extend further north. Peculiar to this region are only six genera of Rodents, which seems an insufficient reason for raising the Sonoran province to the dignity of a region. Considered from the point of view of numbers of peculiar forms, the Thibetan subregion has more claims to distinction as a region; for confined to that area we have the genera Nectogale, Aeluropus, Eupetaurus, Pantholops, Budorcas; while by slightly extending its limits, a number of other peculiar forms might be added. Madagascar has distinctly more claims to regional division. Absolutely confined to it are eleven of the seventeen existing genera of Lemurs, the family Centetidae among the Insectivora, which contains seven genera, and another recently discovered and peculiar genus, Geogale; it has six peculiar genera of Viverridae; it has five peculiar genera of Rodents. In addition to this it is negatively characterised by the absence of the following typical African animals, Felidae, Proboscidea, Rhinocerotidae, Equidae, Monkeys, etc. It seems to be impossible to avoid allowing the rank of a region to this part of the world.

In separating the Nearctic from the Palaearctic region, stress must be laid rather upon the absence of Asiatic and European forms from North America than upon the existence in the northern half of the New World of many peculiar forms. Peculiar to the Nearctic are the Goat genus Haploceros, the Rodents Erethizon, Zapus, and the family Haplodontidae. The Mole genus Condylura is also restricted to this part of the New World. Even so it has more peculiar forms than the Sonoran. If we add to this the absence of Horses, Antelopes except Antilocapra, Pigs, Hyaenas, etc., there are strong grounds for retaining this division. It must be agreed, however, that it comes rather nearer to the Eurasian district than the latter does to the Oriental.

The Oriental region has many characteristic animals. It has among the Anthropoid Apes the Orangs and Gibbons; of Old World Apes it has confined to its own area the genera Semnopithecus and Nasalis. Of Lemurs there are Loris and Nycticebus, and Tarsius, representing a family of that order, or even a sub-order. The Galeopithecidae are entirely Malayan. There are many Rodent, Carnivorous, and Insectivorous genera; the Rhinoceroses and the Elephant of this region differ from those of Africa.

Tragulus concludes a sample from a very rich list of peculiar forms.

The Ethiopian region has also its Anthropoids, the Gorilla and the Chimpanzee, but they belong to genera or a genus different from those which include the Oriental forms. There are five peculiar genera of Cercopithecidae. The Lemurs restricted to this region are Galago, Perodicticus and Arctocebus. The peculiar Insectivorous families Macroscelidae and Chrysochloridae are only found here, besides many other peculiar genera. Africa is especially the home of Antelopes, and the Giraffe is not found now outside its borders. The Elephant and the Rhinoceroses are of different species from those of India. There are many peculiar Rodents and Ungulates.

CHAPTER III

THE POSSIBLE FORERUNNERS OF THE MAMMALIA

The relationship of Mammals to Vertebrates lying below them in the scale, their origin in fact, is a much-debated question, with many attempted solutions. To enter into this large question in detail would involve a great deal of useless statement of arguments founded upon misleading or upon quite inaccurate "facts." It will perhaps be sufficient if we reflect here the current view most in vogue at the present, i.e. that which would refer the Mammalia to reptiles belonging to the extinct Permian and Triassic group of the Theromorpha (also called Anomodontia). These have been explored lately to a very large extent, and chiefly by Professor Seeley.[^[43]] The very fact that a genus Tritylodon, only known by the forepart of the skull, has been called Mammalian and Anomodont by various authors, shows at least the difficulty of differentiating the two groups when the material for study is imperfect. As a matter of fact these Theromorpha are without doubt reptiles; they show, for example, a lower jaw formed out of several distinct pieces, of which the articular articulates with a fixed quadrate on the skull. They possess the characteristic reptilian bones, the "transverse," the pre- and post-frontals, and there are various other points of structure which leave no room for doubt as to their truly reptilian nature. There are, however, numerous indications of an evolution in the mammalian direction in all parts of the skeleton, to the more important of which some reference will be made here. It may be as well to clear the

ground by mentioning the fact that among the Theromorpha four distinct types of reptiles are included, which are considered to form four orders, i.e. the Pareiasauri, the Theriodontia, the Anomodontia (Dicynodontia), and the Placodontia.

The first of these divisions includes what seem to be basal forms. These reptiles show numerous points of likeness to the Amphibian Labyrinthodonts.[^[44]] On the other hand the third division, that of the Dicynodontia, are highly-specialised Theromorpha, from which no further evolution would appear to have been possible. Thus the dentition was either completely lost, or reduced to tusks as in Dicynodon. We need not therefore concern ourselves in the present volume with these Anomodonts. It is with the Theriodonts that our business lies. The very name, be it observed, is aptly chosen on the hypothesis to be explained here; but it is not only in the teeth that these reptiles show likenesses to the Theria or Mammals, but in almost every feature of their organisation. Unlike other reptiles, the Theromorpha in general were lifted comparatively high above the ground on legs of fair length and of mammalian relationship in the position of the segments of the limbs. The typical reptile grovels upon the earth with legs sprawling out, as indeed the very name suggests. One bar to the Theriodonts being on the direct line of mammalian ancestry has been urged as a preliminary difficulty, and that is their large size. The earliest undoubted mammals were small creatures, comparable to a Rat or a Mouse in size; whereas a good-sized Bear or a Wolf is a better standard of size for some of the best-known genera of Theriodonts. It has, however, been quite permissibly suggested that living in company with these large Theriodonts were less obtrusive genera, from which the mammals might have sprung. It is so familiar a fact that a given group of animals generally contains giants, dwarfs, and members of intermediate size, that this suggestion may almost be accepted as a fact. It need at least present no difficulties to us in our comparisons.

The most salient "mammalian" feature of the Theriodonts is the heterodonty of the teeth, the pattern of the "molars," and the limited number which constitute the series. The fact, too, that they are limited to the dentary bones below and to the

maxillae and the premaxillae above, is a sine qua non for mammalian comparison. In the more basal Theromorpha the teeth are not so limited in position. Finally, to complete the remarkable mammalian resemblance of the teeth of these reptiles, it must be mentioned that in Tritylodon and Diademodon the roots of the molars, as we may fairly term them, though not actually divided after the mammalian fashion, were deeply marked by a groove, which suggests an incipient division or a fusion of two distinct roots. Some of these facts of structure may now be considered in further detail. As to the incisors and canines, it is sufficient to say that the numbers of the former, and the shape of the latter, are in perfect consonance with a derivation of the Mammalia from this group. The molar series can be divided into premolars and molars, at least in so far as regards their shape; for the anterior teeth are often smaller and less complicated than those which follow, as is often the case with the two series in mammals. The molar series also consist of teeth in close apposition to each other and separated from the canines by a diastema, which is a character of mammalian teeth. The fact that in the reptile Cynognathus and the mammal Myrmecobius there are nine of these molar teeth in each half of each jaw is perhaps not a point upon which it is desirable to dwell with too much weight; but the general fact that the molars are further reduced in some genera of Theriodontia than in that which has been mentioned, is clearly a matter of significance when the ancestry of the mammals is under consideration.

The most interesting fact about the molar series in the Theriodontia is that we meet with the two types of molars that occur in the mammals. Cynognathus and other genera have molars which consist of a main cusp, and of one cusp before and one after the main cusp; in fact these teeth are triconodont as in certain early mammals, a state of affairs which is believed by the "trituberculists" (see p. [56]) to have preceded the tritubercular tooth. There are also "multitubercular" teeth, especially well developed in Tritylodon, where they exactly resemble those of certain Multituberculata, and whose structure originally led to the placing of Tritylodon among the mammals of that group. If there is any question about the mammalian nature of this fossil, there remain several other Theriodontia in which the multituberculism is well marked. It is so in Trirhackodon

and in Diademodon for instance. This incidentally lends some support to the idea that the Mammalia have been evolved from two sources, a way of looking at the origin of the group that will coincide with the views of some authors like the late Dr. Mivart, and will at the same time reconcile the trituberculists and the multituberculists. For we should then assume that the Eutheria and Triconodontia had originated from some such form as Cynognathus; and the Multituberculata and the existing Monotremes from some form like Diademodon. It is not of great use to point out that Diademodon is really of the trituberculate pattern, because in its molars, though multituberculate, the trituberculate main cones can be recognised; for that state of affairs could just as well have been brought about by a reduction from the multituberculate type. The skull of these Theriodonts shows some well-marked approximations to the mammalian type. There is in the first place a commencing consolidation and reduction of the individual bones, which is so distinguishing a feature of the mammalian skull as opposed to the skull of lower vertebrates. In Cynognathus the postorbital is fused with the jugal, and the supratemporal with the squamosal, forming apparently one bone. In the lower jaw the splenial is often reduced to the thinness of paper, thus indicating a commencing disappearance. In many Theromorpha the squamosal shares largely in the formation of the articular facet for the lower jaw, obviously an important mammalian characteristic; this is brought about by the reduction of the quadrate, which latter bone, moreover, acquires in certain particulars the appearance of the mammalian malleus, with which it is, according to many, homologous. But this subject has been already dealt with on page [26]. A very pronounced likeness to the mammalian skull is that there are two occipital condyles. That this has been brought about by the further development of a tripartite condyle such as occurs in tortoises, and that by the suppression of the basi-occipital part, does not affect the resemblance to the mammalian skull; in fact it explains the origin of two condyles from the typical reptilian single condyle, and disposes of the necessity for believing, with Huxley and others, the Amphibia to be on the main line of mammalian evolution on account of their two condyles. The general aspect of the skull in Cynognathus has been

compared to that "of Thylacinus or Dissacus." No one can examine the actual sketches of the skull of that Theriodont without endorsing that opinion. As a curious detailed point of likeness to certain Mammalia may be mentioned "a small descending process of the malar bone, which may be a diminutive representative of the descending element of the malar seen in Elotherium, Nototherium, Diprotodon, Macropus, certain Edentata, such as Glyptodon, Megatherium, Mylodon, Bradypus, but unparalleled so far as I am aware in fossil reptiles." (Osborn.) The zoologist cannot help being impressed with the significance of small details of similarity, which do not seem to be due in any way to surrounding conditions of life, and thus referable to mere convergence, like the fish-like form of Whales and Seals.

The rest of the skeleton of the Theriodontia is by no means so well known as the skull and teeth. But from what is known, other mammalian characters can be pointed out. Perhaps the most striking mammalian feature is to be found in the scapula of Cynognathus. It is in this creature somewhat narrow and elongated; but it has a well-marked spine, ending in a hooked acromion. Now it is to be noted in support, so far, of the diphyletic origin of mammals, that in the Monotreme, as in Whales indeed, the spine forms the anterior border of the scapula, and is coincident with it, there being thus no prescapula at all in the Monotreme, and only a trace of it in certain Whales.[^[45]] Whether the multituberculate Tritylodon or Diademodon had a scapula after the Monotreme pattern is not known; but it is clear that the scapula of the triconodont Cynognathus is quite after the pattern of the Eutherian scapula. Furthermore, Professor Seeley is of opinion that the coracoid was relatively small, and indeed smaller than the same bone in Edentates, and a fortiori than in Monotremes. Another fact of structure which points also, possibly, in the direction of a diphyletic origin for the Mammalia, is the double-headed ribs of Cynognathus. As is well known, the ribs of the Monotremata have only the central head, the capitulum.

As a general mark of affinity with mammals the reduction of the intercentra in Cynognathus may be noted, and also the existence of a small though perfectly obvious obturator-foramen, separating the pubis from the ischium. There are further details

which tend in the same direction. And we shall probably not go far wrong in the present state of our knowledge if we assign the origin of the mammals to some type which would be included in the order Theriodontia or at least in the sub-class Theromorpha.

CHAPTER IV

THE DAWN OF MAMMALIAN LIFE

The animals that we considered in the last chapter, though showing certain unmistakable likenesses to the mammals, are nevertheless unquestionably not mammals but reptiles. In the Triassic strata, however, we first meet with the remains of undoubted mammals. The Mammalia first appeared upon the earth in a tentative and hesitating way: they had not cast off many of the characters of their supposed reptilian forefathers; they shrank from observation and destruction by their small size, and apparently, so far at any rate as their teeth afford a clue, by an omnivorous diet. The world abounded at that period in large and carnivorous reptiles, which may indeed have been the principal enemies with which the first mammals had to cope. These early mammals lingered on to so late a period as the Eocene; but the majority of the genera were Triassic, Jurassic, and Cretaceous. Certain of the primitive mammalian forms have been referred to the Marsupials, and their resemblances to the Monotremata have also been pointed out. The current view of the present time, however, is that they form a special order, which may possibly have embraced the ancestors of both Marsupials and Monotremes; for it is reasonable to explain in this way the combination of characters of these two orders which they present. For this group the name Allotheria has been proposed by Marsh, and Multituberculata by Cope; the latter term is the less suitable, in that the Monotremata (Ornithorhynchus) are also "multituberculate." The group is known in a very imperfect fashion. The remains are but few and fragmentary; and for the most part we have only a few teeth to speculate upon. This is natural enough, for the harder teeth might easily be supposed to

have resisted the decay which would more readily affect the softer bones. Where there are bones it is frequently the lower jaw alone which has been preserved for us—a bone which has also been preserved in the case of some of the contemporary Marsupials.

It has been pointed out (from the observation of dead dogs floating in canals) that the lower jaw is occasionally detached from the carcase. It is the most readily separable part which contains a skeleton. It may be, therefore, that the remains of these early mammals, floating down some river to the sea, may have lost their jaws while in the river, or at furthest in the shallow waters of the sea, the rest of the carcase floating out to a greater distance, and being finally entombed in the stomach of some carnivorous fish, or in the mud at the bottom of a deep ocean, which has never since seen the light.

The characters of this group are really more those of the Monotremata than of the Marsupials. The undoubted likeness which their molar teeth show to the temporary teeth of the Platypus have already been commented upon. Like the Monotremes the Allotheria appear to have possessed a large and independent coracoid; the evidence for this rests upon the discovery of the lower end of a scapula of Camptomus, a Cretaceous genus from North America upon which there is a distinct facet for the articulation of what can have been nothing else than a coracoid. On the other hand they differ from the Monotremata by the presence of incisor teeth which were Rodent-like in form, and not very different from those of certain Marsupials. This point of difference cannot be regarded as of very first-rate importance; no one would relegate the Sloth and the Armadillo to different orders on account of their tooth differences, which are about on a par with those to which we have just referred. It seems indeed likely that it will be ultimately necessary to rub out the boundary line which now divides the Allotheria and the Monotremata.

The Plagiaulacidae are unquestionably mammals, and they are placed by most naturalists in this at present uncertain group of Multituberculata, which will be retained here in deference to the distinguished authorities who have instituted the group, though there are but few characters by which it can be defined. This family though appearing in the Trias, extends down in time to the Eocene. The type-genus, that which has given its name to

the family, is Plagiaulax. As it is not Triassic, the consideration of its characters will be deferred until later. Microlestes is a Rhaetic genus, known from rocks in Germany and England; but it is entirely based upon molar teeth. M. antiquus has a two-rooted molar of an elongated form with a row of tubercles on either side of a median groove, which traverses the long axis of the tooth. To some extent the teeth of the ancient form resemble those of Ornithorhynchus. Microlestes has been sometimes spoken of as a Marsupial, but Mr. Tomes[^[46]] has found that it does not show one very universal character of the Marsupial teeth: it has not those continuations of the dentinal tubes which traverse the enamel in all Marsupials that have been examined with the sole exception of the Wombat.

The rarity of the remains of mammals in these earliest rocks of the Secondary epoch has been accounted for in another way from that which has been suggested above. It may be that the group Mammalia was not evolved in Europe at all, and that the stray remains which have been found in that continent represent the fragmentary remnants of a few scattered immigrants which heralded the later invasion of more numerous genera during the Jurassic period.

The Mammals of the Jurassic Period.—Some of the Allotheria or Multituberculata described in the last section occur in the rocks of this early part of the Secondary epoch. They are doubtful in position, as already stated; some of them indeed, as for instance Tritylodon and Dromatherium, are possibly not mammals at all, while the remainder probably belong to a non-existent order of mammals. Along with these dubious creatures are the fragmentary remains of small animals which are not merely mammals, but in all probability definitely Marsupials. It is true that here again we have little beyond lower jaws and teeth to deal with; so that there may be less certainty in referring them to the Marsupials than appears to be the opinion of the majority of Palaeontologists.

Professor Osborn in fact considers that the Mesozoic mammals consist of three groups: (1) The Multituberculata, including the Bolodontidae, Stereognathidae, Plagiaulacidae, Polymastodontidae, and possibly the Tritylodontidae (which, however, are regarded by him and by others as more probably reptiles of the

Theromorphous group). (2) The Triconodonta, which were Marsupials, though in all probability with a complete succession of teeth and with an allantoic placentation. This group will include the genera Phascolotherium and Amphilestes, as well as Triconodon and Spalacotherium. Finally we have (3) the Trituberculata (or Insectivora Primitiva) with the genera Amphitherium, Peramus, Amblotherium, Stylacodon, and Dryolestes.

We shall take these three groups in order. The Multituberculata have already been to some extent defined, if such a word can be used to express the summation of the very scanty information at our disposal. Of this group, Plagiaulax is a genus which occurs in the Purbeck beds; it is only known by lower jaws implying an animal of the size of a Rat or rather smaller. The jaws have in front a large incisor which looks Rodent-like, and also like those of the Diprotodont Marsupials; but it is held that these teeth did not grow from persistent pulps, and there is in any case no anterior thickened coating of enamel. Canines are absent; the diastema is followed by four premolars increasing progressively in size and possessing somewhat complicated grinding surfaces. These surfaces are formed by several obliquely-set ridges. The succeeding teeth are termed molars on account of their difference in structure, and there are but two of them on each side. The molars are of a pattern common in the Multituberculata; the centre is hollowed, and the raised rim is beset by tubercles. Other Jurassic genera of Multituberculates are Bolodon, Allodon, and Stereognathus. All of these possess the same multituberculate molars.

Of the Triconodonta the type-genus is Triconodon. This genus is better known than most Jurassic mammals, since both the upper and the lower dentition have been described. It appears to have possessed the typical Eutherian dentition of forty-four teeth, to which a fourth molar is added in some species. The great difference between the molars and premolars argues a complete tooth-change. The genus is American as well as European.

Spalacotherium has more molars, five or six.

Phascolotherium bucklandi, on the other hand, is a much older type in the form of its teeth. There are, however, not so many of them as in Amphitherium; Phascolotherium has but two premolars and five molars, making a total of forty-eight teeth. The teeth are of the triconodont form, the three cusps being in line, and the middle one the largest.

Amphilestes has teeth of the same pattern but has more of them, the premolars and molars being respectively four and five. All these animals had the lower jaw inflected. Whether they are all Marsupials or not, it is clear that Phascolotherium and Amphilestes should be united and placed away from Amphitherium on account of the more primitive form of their teeth.

We next come to the Trituberculata.

Among the most celebrated of these remains are a few jaws discovered in the Stonesfield slates near Oxford, and examined by Buckland, Cuvier, and some of the most eminent naturalists of the beginning of the last century. These jaws have been lately submitted to a careful re-examination by Mr. Goodrich,[^[47]] who has increased our knowledge of the subject by exposing from the rocky matrix in which the jaws lie fresh details of their structure; it is probable therefore that now all that there is to be learnt from these specimens has been recorded.

Amphitherium prevostii was a creature about the size of a Rat. Its jaw was first brought to Dean Buckland about the year 1814, and described six years later. Buckland thought the jaw to be that of an Opossum, an opinion in which Cuvier concurred. The jaw, however, is marked by a groove running along its length, and this groove was regarded by de Blainville as evidence of the composition of the jaw out of more than one element, which would naturally lead to its being regarded as the jaw of a reptile.[^[48]] This species and another named after Sir Richard Owen have a dental formula which, like that of the Marsupials, is large as compared with that of the Placental mammals; it runs: I 4, C 1, Pm 5, M 6—i.e. 64 teeth altogether. This is a larger number than we find in any existing Marsupial. But as in Marsupials, and in certain Insectivora also, the angle of the jaw is inflected. These teeth are of the tritubercular pattern with a "heel." They are in fact closely like those of the living Myrmecobius; but not, it should be remarked, unlike those of certain Insectivora.

The Mammals of the Cretaceous Period.—At one time there was a totally inexplicable gap between the Jurassic and the basal Eocene, a series of strata which occupy an enormous expanse of time in the history of the earth having appeared to

be devoid of mammalian remains. This gap, however, has been filled up by the discovery of mammalian remains in the North American Laramic formation, which seems to be clearly of Cretaceous age. Furthermore, it is held by some that the Purbeck beds are more properly to be placed with the Cretaceous, which would then necessitate the consideration under the present heading of some of the types already dealt with; and if, as is suggested in the following section, the lowest so-called Eocene beds are really referable to the Cretaceous, there is no lack of mammalian remains in that period. And, moreover, it was in that case the Cretaceous period which witnessed the evolution of the existing orders of Placental mammals. Otherwise the mammalian remains of the Cretaceous agree with those of the Jurassic. We find remains of the Multituberculata in fragments of Plagiaulacidae and Polymastodontidae. Ptilodus is a genus which has two premolars; and Meniscoessus is another multituberculate from the same Laramic formation. The other detached fragments of mammals are thought by Osborn to represent both Placentals and Marsupials.

The Mammals of the Tertiary Period.—Unless the lowest beds of the earliest Tertiary period, the Eocene, such as the Torrejon of North America, should be in reality referred to the Cretaceous, there is no evidence that the modern groups of Mammalia existed before the present epoch of the earth's history. It is probable, however, that the Eutheria as a group were Mesozoic. The fossil jaws that have been considered in the last chapter may quite probably be primitive Eutherians, or even divisible, as believed by Professor Osborn, into Marsupials and Insectivores. In the Tertiary, however, apart from the question as to the nature of the Puerco and Torrejon formations, and as to certain South American strata whose fossil contents have been investigated by Professor Ameghino, we find the first traces of mammals definitely referable to existing orders, or to be distinctly compared with existing orders. Since, however, representatives of types which have obvious relationships to modern types appear in considerable profusion in the very earliest strata of the Eocene, it seems clear that much remains to be discovered in beds earlier than these. Confining ourselves, however, to facts and to comparisons which can be made on more than a few lower jaws and scattered teeth, which is practically all that we

possess of earlier mammals, we must arrive at the general conclusion that two of the existing larger groups of the Eutherian, non-Marsupial, mammals were differentiated at quite the beginning of the Eocene, and were represented by forms from which it is possible to derive at least the existing Carnivora, Insectivora, Artiodactyla, and Perissodactyla. These were the Creodonta and the Ungulate Condylarthra. In addition to these we may enumerate as very early types the Lemuroidea, represented by such forms as Indrodon in the New World, and (though later) by Necrolemur, etc., in the Old World, and the Edentata, if we are to allow as their ancestors the Ganodonta.

The early Eocene strata also contain representatives of at least one order, the Amblypoda, which increased subsequently, but has died out without descendants, unless we are to believe with some that the Elephants are to be derived from these Eocene "pachyderms." In later Eocene times the great majority of the existing orders, and even subdivisions of orders, are to be met with; and there are in addition such totally extinct orders as the Typotheria, Ancylopoda, and Tillodontia. Coupled with this gradual specialisation in the orders of Eutherian mammals, there is naturally a vast increase in the number of generic and family types. This culminates perhaps in the Miocene, from which time there has been a gradual decline in mammalian variety, so that it is justly said that we live now in an epoch which is impoverished of mammals. This gradual decay has persisted until to-day, as is witnessed by the extinction of the Rhytina and the Quagga, and the growing rarity of the White Rhinoceros and the American Bison.

The early Eutherian stock consisted of small mammals with small heads and slender, long tails. The limbs were pentadactyle, ensheathed in claws or broader hoofs. The fore-limbs may have been partly prehensile. The teeth were forty-four, completely differentiated into incisors, canines, molars, and premolars; and there appears to have been a complete diphyodontism. The canines were not greatly enlarged, and no diastema separated any of the teeth. The molars were bunodont or of a more cutting pattern, with some five or six tubercles. These animals were, moreover, very small-brained. This early stock is represented by Creodont and Condylarthrous animals, the exact boundaries between which are hardly marked in the

very early types. Professor Osborn has argued that from this early Eutherian stock there were two waves of progress, or, as he expresses it, "two great centres of functional radiation."[^[49]]

The first was largely ineffective, the second has produced all the Eutherian orders of to-day. These two divisions are termed by him "Mesoplacentalia" and "Cenoplacentalia." The first division embraces the Amblypoda and their descendants the Coryphodonts and Dinocerata, many of the Condylarthra, the bulk of the Creodonts and the Tillodonts. These creatures persisted for a time, but died out in the Miocene. They were mainly distinguished by the smallness of their brain; the great specialisation of structure which they exhibit having left that organ unaffected, and therefore tending in the long run to render them unable to cope with changes in the inorganic and organic world. The successful division of the primitive Eutheria comprises the groups which exist at the present day, and is not connected directly with those small-brained Mesoplacentals; it has apparently originated, however, from the least specialised of their ancestors. Professor Osborn thinks, moreover, that the Lemurs and the Insectivores are persistent descendants of the earlier wave of Eutherian life. It appears in fact as if Nature had created the existing Ungulate, Unguiculate, and other types on a defective plan, and, instead of mending them to suit more modern requirements, had evolved an entirely new set of similarly-organised types from some of the more ancient and plastic forms remaining over. The Marsupials may be the only group of the early wave remaining, and they have been able to hold their own for the geological reason that Australia was early cut off from communication with the rest of the world. That they are disappearing seems to be shown by their gradual diminution as we pass from Australia towards the continent of Asia, through the islands of the Malay Archipelago. Competition has here decimated them, as it may do in the remote future in Australia.

It is often said, but with some looseness of statement, that ancient quadrupeds are huger than their modern representatives. This statement is partly true in fact, but largely wrong in implication. For it suggests that—and the suggestion is often expressed in books that are not authoritative—huge animals

have left a dwarfish offspring; that there were giants of old, and that there is a puny race to-day. As a matter of fact, the study of the gradual evolution of the early Tertiary Mammalia into their descendants of later times shows very plainly the truth of this interesting generalisation: That the primitive types were all small creatures, and that in those instances where we can trace a pedigree, there was a gradual increase in size up to a point where greater increase led to extinction. We point out later on a number of facts illustrating this matter in detail. It has been ascertained, for instance, that the pedigree of the Horses, the Camels, the Rhinoceroses, and many other groups, commences with small forms and culminates in large ones. It may be urged that such animals as the Tapir are to-day smallish forms, and that related to them in the past were the gigantic Titanotheres; but in this and similar cases it will be found that the extinct giants were not in the direct line of pedigree, but represented side-branches which waxed huge on their own account and then disappeared.

CHAPTER V

THE EXISTING ORDERS OF MAMMALS

Prototheria—Monotremata

Apart from those creatures whose fragmentary remains have been considered in the last chapter, and which belong to the earliest of mammaliferous strata, the remains of Mammalia are all referable to existing orders. In the pages which follow we shall therefore deal with the actual representatives of living families side by side with their extinct relatives. The existing orders of Mammalia, together with those of their fossil allies, can be plainly divided into two great subdivisions, or, as we shall term them, sub-classes; the Mammalia as a whole being termed a class of the Vertebrata comparable with the class Reptilia, etc. It has been usual, owing to the initiative of Professor Huxley, to divide the Mammalia into three divisions of primary importance. We shall adduce reasons later for not accepting this mode of division, but that which allows of only two primary divisions. These two divisions are (1) Prototheria and (2) Eutheria. Whether the Multituberculata, Trituberculata, and Triconodonta, considered in the last chapter, are really to be distributed among these two sub-classes is a matter upon which it is possible to form an opinion, but not to dogmatise. The Prototheria stand at the base of the mammalian series, and present many likenesses to the Sauropsida; the Eutheria are the animals which are most fully differentiated as mammals. We shall commence with

Sub-Class I.—PROTOTHERIA.

To this group belongs the order Monotremata, and possibly also the so-called Allotheria or Multituberculata. As, however,

the latter are only known from very fragmentary remains, which are not sufficient to determine the systematic position of the animals of which they are fragments, I have not thought it worth while to attempt a serious definition of the order Multituberculata. I have introduced a short account of the principal facts which are known concerning the creatures grouped together under this name into the historical sketch of the progress of mammalian life in Chapter IV. As to the Monotremata, there is no question that they are entitled to rank in a group equivalent to that including all other mammals of which we have sufficient knowledge to construct a classificatory scheme. There have been, indeed, naturalists, such as Meckel, who would altogether deny the mammalian rank of these creatures.

The Monotremata or Ornithodelphia may be thus defined:—

Mammalia with no teats, but with a temporary pouch in which the young are hatched, or to which they are transferred after hatching, and into which open the ducts of the mammary glands. An anterior abdominal vein, or at least the membrane supporting it, persists throughout the abdominal cavity. Heart with an incomplete and largely fleshy right auriculo-ventricular valve. Brain without a corpus callosum. Shoulder girdle with a large coracoid reaching the sternum; clavicles and an interclavicle present. There are "marsupial" or epipubic bones attached to the pelvis. Vertebrae with no epiphyses for the most part. Ribs with only capitulum and no tuberculum. Mammary glands of the sudoriparous and not the sebaceous type of epidermic gland.[^[50]] Oviparous, with a large-yolked and meroblastic ovum, enclosed within a follicle of two rows of cells.

To call these animals Mammalia is of course an abuse of the meaning of that word in one sense, but it is not in another; since the pouch of these Monotremes is, as has been explained elsewhere (p. [16]), the real equivalent of a teat, and not of the pouch of the Marsupials.

The most salient characteristic of this group of mammals in the estimation of their position in the vertebrate series is not so much the fact that they are oviparous as that the eggs are large-yolked, and develop therefore, so far as regards their early stages, after the fashion of the egg of a reptile. The laying

of eggs, or at least ovoviviparity, would follow from the structure of the egg, since the abundance of yolk would do away with the necessity for a placenta. That the eggs had this Saurian characteristic was first definitely made known by Professor Poulton[^[51]] for Ornithorhynchus, and his results were confirmed later for Echidna.[^[52]] The structure of the eggs has, however, already been dealt with on p. [72]. The fact that these animals lay eggs appears to have been known for a very long time, though rediscovered so lately as 1884 by Mr. Caldwell.[^[53]] In connexion with the structure of the ova, the ovaries themselves and the oviducts are built upon the Sauropsidan plan. In the male the testes retain the primitive abdominal position. The fact that the urinary and genital products escape by means of their ducts into a chamber which also receives the end of the alimentary tract is not a distinctive feature of this group, inasmuch as it is seen in the Marsupials, and also in certain low Eutheria, such as the Beaver and other Rodents, and a few Insectivores. As to external features, the Monotremata show certain archaic characters. The unspecialised arrangement of the mammary glands has already been described. These animals are plantigrade, if the term may be used also to describe the aquatic Ornithorhynchus. The ears are absolutely destitute of a conch. The remarkable spur upon the hind-legs furnished with a gland, which is more marked in the male, and indeed disappears in the female of Ornithorhynchus, is a structure which argues the specialised condition of these two modern representatives of what must have been a large order in the past.

Fig. 51.—Ventral view of skull of Echidna aculeata, and right half of mandible, ang, Angle of mandible; aud.oss, auditory ossicles; cond, condyle of mandible; cor, coronoid process; max, maxilla; oc.cond, occipital condyle; pal, palatine; p.max, premaxilla; pt, pterygoid; sq, squamosal; ty, tympanic ring. (After Parker and Haswell.)

The skeleton shows numerous ancient characteristics. In the skull there is no demarcation of the orbit from the temporal fossa, a feature widely found in archaic mammals. The tympanic remains as a slender ring, there being no auditory bulla formed either from this or from any other bone. The malleus and incus are large, and thus reminiscent of the quadrate and articular bone of reptiles. In the lower jaw the absence of a marked coronoid process, and the absence of a firm ossification at the meeting of the two rami, may be a primitive state of affairs. It must be remembered, however, that the Cetacea show the same characters, though it is possible that they too are developed from a low mammalian stock. In the vertebral column we find the typical mammalian seven cervicals; but those characteristically mammalian structures the epiphyses are totally absent in Echidna, and only to be seen in the tail-region in Ornithorhynchus. In having only the capitular head to the ribs, these mammals are evidently far removed from all other mammals, and are even more reptilian than the Theromorphous reptiles. The large clavicles and the interclavicle (Fig. 52,

p. 109) are characteristic of the group, and the latter bone is peculiar to the Monotremata among mammals. So, too, is the large coracoid. In the scapula there is a spine which coincides with the anterior border of that bone. The arrangement of the muscles in this region proves conclusively that this projection is the homologue of the spine and the acromion of other mammals. Here, again, we have a point of likeness to the Cetacea.[^[54]] In the pelvis the acetabulum is perforate (in Echidna), as in Sauropsida.

Fig. 52.—Side view of right half of the shoulder girdle of a young Echidna (Echidna aculeata). × 1. a, Acromion; c, coracoid; cb, coracoid border; cl, clavicle; css, coraco-scapular suture; ec, epicoracoid; gb, glenoid border; gc, glenoid cavity; ic, interclavicle; pf, postscapular fossa; ps, presternum; s, spine; ss, suprascapular epiphysis; ssf, subscapular fossa. (From Flower's Osteology.)

Considering the numerous very archaic features which the general structure of this group displays, it is surprising to find how typically mammalian they are in certain other peculiarities. The mammalian diaphragm, one of the distinguishing features of the class, is perfectly normal in the Monotremata. The alimentary canal shows no great divergences from the normal structure. The stomach is almost globular, with a projecting pyloric region in Ornithorhynchus; the intestine is divided into a "small" and "large" intestine by a slender caecum. The liver has the subdivisions that this organ usually shows in the Mammalia. However, the presence of the ventral mesentery and of the abdominal vein in Echidna and Ornithorhynchus has already been mentioned as a distinctive character. The peculiar and apparently partly primitive valve of the right ventricle has been described above (see p. [66]). The brain is in most respects mammalian in its characters, but naturally shows some important differences. Dr. Elliot Smith, who has most recently studied this question,[^[55]] is of opinion that the size of the cerebral hemispheres is not at all reptilian; indeed, it "greatly exceeds that of

many other mammals." In Echidna, too, but not in Ornithorhynchus, the hemispheres are well convoluted, though the arrangement of these convolutions cannot be brought into line with what is known concerning the convolutions upon the hemispheres of other mammals. It had been stated that in these animals, at least in Echidna, there were only two optic lobes, as in lower vertebrates, instead of the mammalian four. The late Sir W. H. Flower set this matter at rest,[^[56]] and showed that Echidna was in this respect typically mammalian. The absence of the corpus callosum is one of the principal features separating the Monotremes from other mammals.

The Monotremata are represented to-day by two types, Ornithorhynchus and Echidna, which are no doubt worthy of being placed in separate families. Fossil remains of the group (apart from the problematical Multituberculata) are only known from Pleistocene times in Australia, and consist of the bones of a large species of Echidna, and some fragments of Ornithorhynchus, indicating a smaller animal than the living Platypus.

Fig. 53.—Brain of Echidna aculeata, dorsal view. (Nat. size.) (From Parker and Haswell's Zoology.)

Fam. 1. Echidnidae.—This family contains two genera, of which Echidna is the older and much the better known. The skin is abundantly covered with spines, with which are mingled hairs. The snout is tapering, the tail rudimentary, and the fingers and toes five in number. The spur and gland upon the calcaneum are smaller than in Ornithorhynchus. The claws are very strong, serving to tear open the ants' nests, upon the inhabitants of which the Echidna feeds, licking them up with a long extensile tongue like that of Myrmecophaga. In relation to this habit the salivary glands are enormously developed, and indeed the animal has been confounded with Myrmecophaga,[^[57]] as the vernacular name "Australian Anteater" exemplifies.

In the skull the Echidna differs from Ornithorhynchus in the greater extension backwards of the palatines, and the larger size of the pterygoids. The extent and relations of these bones to each other is not at all unlike that which obtains in many Whales. The premaxillae show traces of the same divergence followed by convergence of their ends that is seen in the Platypus. There are only sixteen pairs of ribs, and either three or four lumbar vertebrae. Echidna has no trace of teeth, and there are no horny pads which take their place; the mouth is as edentulous as in the true American Anteaters. The brain (Fig. 53) is marked by sulci, contrary to what we find in Ornithorhynchus. The genus has been divided into three species, but it is doubtful whether more than one can be allowed, which ranges from Australia through the Papuan region. While there is but one species of true Echidna, a New Guinea species must clearly be referred to a distinct genus Proechidna.[^[58]] This animal is to be distinguished by the fact that there are usually but three toes on each foot. But there are copious rudiments of the other phalanges, upon which claws are sometimes developed. The beak is curved downwards, and the back is rather arched; the whole animal has the most singular likeness to an Elephant! The ribs are increased by one pair, and there are four lumbar vertebrae. The one species is named P. bruijnii. The Hon. W. Rothschild[^[59]] distinguishes a form P. nigroaculeata, which is allowed by Mr. Lydekker.

Fig. 54.—Australian Anteater. Echidna aculeata. × 1⁄6.

The Echidna feeds like anteaters, by thrusting its tongue into an ant-hill, and waiting until it is covered with indignant and marauding ants, which are then swallowed. But this animal also devours worms and insects, which are extracted from their hiding-places by the tongue. It is mainly nocturnal, and prefers the seclusion of the densest scrubs of the bush, or rocky spots where it is free from intrusion. Dr. Semon did not find that the spur of this animal was used at all in self-defence; but he thinks that possibly the weapon may be used, in the breeding season only, in the combats of the males for the females, when perhaps, as has been shown to be the case in Ornithorhynchus, the gland attached to it produces a poisonous secretion.

The egg, as it appears, is transferred to the pouch by the mouth of the mother; the shell is broken by the emerging young one, which has an egg-breaking tubercle on its snout for this purpose; the mother removes the shell. When the young has attained a certain size, the mother removes it from the pouch, but takes it in from time to time to suckle it. When on her nightly rambles the young one is left in a burrow dug for the purpose. Dr. Semon was able, from his own observations, to substantiate this act of intelligence on the part of the Echidna. It is well known that the temperature of the Monotremes is less than that of higher mammals; in addition to this fact Dr. Semon found that the range of variation of temperature in the Echidna was as much as 13 degrees or more. It is thus intermediate between the "poikilothermal" reptiles and the "homoeothermal" mammals.

Fam. 2. Ornithorhynchidae.—There is no need to attempt to define this family, since it contains but one genus Ornithorhynchus, with but one species, O. anatinus. The general aspect of the animal is well known. It is covered with dense fur of a blackish brown colour; the limbs are short and five-toed, the toes being webbed. The tail is longish and broad, being flattened from above downwards. The webbing on the anterior toes considerably outdistances the tips of the claws, as in the Seals. But this is not the case with the hind-feet. The "beak," which is broad and flat, and does actually suggest that of a duck, is not covered with horn, as is often stated, but with a fine, soft, sensitive, naked skin, which abounds in sense-organs of a tactile nature. As to characters derived from the skeleton, Ornithorhynchus has

seventeen pairs of ribs and only two lumbar vertebrae. The skull is expanded in front, and the bill is supported by two, at first diverging, and then converging, premaxillae. Between them is the famous "dumb-bell shaped bone," which is believed to be the representative of the reptilian prevomer. The pterygoids are smaller than in Echidna, and the hard palate does not extend so far back as in that genus. The brain of this genus is smooth.

Fig. 55.—Duck-billed Platypus. Ornithorhynchus anatinus. × 1⁄6.

The discovery of the real teeth of Ornithorhynchus only dates from the year 1888, when they were found by Professor Poulton[^[60]] in an embryo. Later Mr. Thomas found[^[61]] that the teeth persist for a considerable portion of the animal's life, and are only shed, like milk teeth, "after being worn down by friction with food and sand." We have already (p. [98]) called attention to the general similarity of these teeth to those of certain of the earliest Mammalia and of mammal-like reptiles. The teeth are all molars, and they are either eight or ten in number. They are replaced by the horny plates of the adult animal; but the mode of replacement is curious. The plates are developed from the epithelium of the mouth, but round and under the true teeth; the epithelium of the mouth grows gradually under the calcified teeth, a method of growth which has possibly something to do with the shedding of the latter. The hollows and

grooves in the plates are the remains of the original alveoli of the teeth.

Fig. 56.—Skeleton of male Ornithorhynchus. Ventral view. The right fore-limb has been separated and turned round so as to bring into view the dorsal surface of the manus. The lower jaw is removed. acc.tars, Accessory tarsal bone supporting the spur; ant.pal.for, anterior palatine foramen; ast, astragalus; atl, atlas; ax, axis; bs.oc, basi-occipital; bs.sph, basi-sphenoid; calc, calcaneum; cbd, cuboid; cerv.rb, cervical rib; clac, clavicle; cond.for, foramen above inner condyle of humerus; cor, coracoid; cun, cuneiform of carpus; dent, horny dental plate; ect.cun, ecto-cuneiform; ent.cun, ento-cuneiform; ep.co, epicoracoid; epist, episternum; ep.pb, epipubis; fb, fibula; fem, femur; for.mag, foramen magnum; glen, glenoid cavity of shoulder-joint; glen, glenoid cavity for mandible; hum, humerus; in.cond, inner condyle of humerus; inf.orb.for, points to position of infra-orbital foramen; infr.proc, inferior processes of caudal vertebrae; int.rbs, intermediate ribs; isch, ischium; mag, magnum of carpus; max, maxilla; max.for, maxillary foramen; metat.I, first metatarsal; metat.V, fifth metatarsal; nas.cart, nasal cartilage; obt, obturator foramen; ol, olecranon; out.cond, outer condyle of humerus; pal, palatine; pat, patella; post.pal.for, posterior palatine foramen; pr.max, premaxilla; pr.st, presternum; pter, pterygoid; pub, pubis; rad, radius; scap, scapula; scaph, scaphoid of tarsus; scaph.lun, scapho-lunar; ses, sesamoid bones of wrist and ankle; sp, tarsal horny spur; sq, squamosal; tib, tibia; trd, trapezoid; trm, trapezium; tym.c, tympanic cavity; uln, ulna; unc, unciform; vom, vomer; x, dumb-bell shaped bone; zyg, zygomatic arch; I-V, digits of manus; V, foramen for fifth nerve. (From Parker's Zoology.)

The Duck-billed Platypus is, as every one knows, an aquatic animal. It is not found all over Australia, but is limited to the southern and eastern parts of that continent, and to Tasmania. The animal excavates a burrow for itself in the bank of the slow streams which it frequents. The burrow has one opening below the water and one above; and it is of some length, twenty to fifty feet. The Platypus feeds upon animal food, chiefly "grubs, worms, snails, and, most of all, mussels." These it stows away when captured into its capacious cheek-pouches. The food is then chewed and swallowed above the surface as the animal drifts slowly along. Dr. Semon, from whose work, In the Australian Bush, this account of the animal's habits is quoted, thinks that in the nature of the food of the creature the explanation of the loss of the teeth is to be found. He is of opinion that for cracking the hard shells of the mollusc Corbicula nepeanensis, upon which Ornithorhynchus mainly feeds, the horny plates are preferable to brittle teeth. Ornithorhynchus is apparently not eaten by the natives by reason of its ancient and fish-like smell. Besides, it is hard to catch on account of its diving capacities, which are aided by an acute sense of sight and of hearing. When the Duck-bill was first brought to this country it was believed to be a deliberate fraud, analogous to the mermaids produced by neatly stitching together the forepart of a monkey and the tail of a salmon.

CHAPTER VI

INTRODUCTION TO THE SUB-CLASS EUTHERIA

Sub-Class II.—EUTHERIA

Definition.—Mammalia with teats. Mammary glands of sebaceous type. Heart with entirely membranous and complete right auriculo-ventricular valve. Brain generally with a corpus callosum. Coracoid much reduced and not reaching sternum. No interclavicle. Vertebrae with epiphyses. Ribs double-headed. Viviparous, with a small ovum.

In this group are included not only the Eutheria in the sense of Huxley, but also his Metatheria. Though the Metatheria, or Marsupials as we shall term them, undoubtedly form a most distinct order of mammals, perhaps even a trifle more distinct than most others, their differences from the remaining tribes are not by any means so great as those which separate Ornithorhynchus and Echidna from all other mammals. In his well-known memoir upon the arrangement of the Mammalia,[^[62]] Professor Huxley enumerated eleven characters as distinguishing the Metatheria either from the Prototheria or from the Eutheria. Of these only three were characters in which they approach the lower mammals. According to his showing, therefore, the preponderance of marsupial features are Eutherian. The three characters of Prototherian type are (1) the presence of epipubes; (2) the small corpus callosum; (3) the absence of an allantoic placenta.

The last of these can be dismissed, in consequence of the recent discovery of an allantoic placenta in Perameles. The first character is apparently a valid distinction between the Marsupials

and their mammalian relatives higher in the series; but it is not a character that should have been made use of by Huxley, since he believed in the existence of a corresponding element in the Dog. As to the corpus callosum (Fig. 50, p. [77]) being small, that seems to be not more than a slight difference of degree.[^[63]] A number of other characters of secondary importance were added by Huxley to the weight of evidence which led him to form a group Metatheria for the Marsupials. Some of these, however, are now known to be not evidence in that direction. For instance he observed that no Marsupial had more than a single successional tooth. It seems at the present moment to be fairly clear that Marsupials have a milk dentition like other Eutherians, but that only one of these teeth, the fourth premolar, comes to functional maturity. That it is really one of a complete milk series is evidenced by the fact that this tooth is differentiated contemporaneously with another series formerly held to belong to the so-called prelacteal dentition.[^[64]] There still remains, of course, the actual fact that the milk dentition is not for the most part functional, but its significance breaks down with these fresh discoveries. Of this Professor Osborn has remarked: "The discovery of the complete double series seems to have removed the last straw from the theory of the marsupial ancestry of the Placentals." But Huxley did not lay much stress upon this matter of the teeth, since he observed that similar suppressions of the milk dentition were to be found in many other mammals admittedly Eutherian.

Fig. 57.—Brain of Echidna aculeata; sagittal section. ant.com, Anterior commissure; cbl, cerebellum; c.mam, corpus mammillare; col.forn, column of the fornix; c.qu, corpora quadrigemina; gang.hab, ganglion habenulare; hip.com, hippocampal commissure; med, medulla oblongata; mid.com, middle commissure; olf, olfactory lobe; opt, optic chiasma; tub.olf, tuberculum olfactorium; vent. 3, third ventricle. (From Parker and Haswell's Zoology.)

Huxley regarded the peculiarities in the reproductive organs

of the Marsupials as "singularly specialised characters," in no way intermediate in character. This view applies also to the pouch, which, as already stated, distinguishes the adults of that group. But the impossibility of using this last character as one of any importance has been shown by the discovery of rudiments of it in embryos of undoubtedly Eutherian mammals (see p. [18]).

Fig. 58.—Sagittal section of brain of Rock Wallaby (Petrogale penicillata). ant.com, Anterior commissure; cbl, cerebellum; c.mam, corpus mammillare; c.qu, corpora quadrigemina; crur, crura cerebri; epi, epiphysis, with the posterior commissure immediately behind it; f.mon, position of foramen of Monro; hip.com, hippocampal commissure, consisting here of two layers continuous behind at the spleneium, somewhat divergent in front where the septum lucidum extends between them; hypo, hypophysis; med, medulla oblongata; mid.com, middle commissure; olf, olfactory lobe; opt, optic chiasma; vent. 3, third ventricle. (From Parker and Haswell's Zoology.)

Less stress is laid now upon the existence of four molars in the Marsupials as dividing them from the higher mammals than was formerly the case. The total dentition of the group is on the whole composed of more numerous individual teeth than in the typical Eutheria; but we have exceptions like the Whales, the Armadillo Priodontes, and the Manatee; or better, because free from the suspicion of secondary multiplication, Otocyon and occasionally (according to Mr. Thomas) Centetes. In the last two there are at least sometimes four molars.

On the other hand, a few archaic characters of some importance crop up here and there among the Marsupials, which are sometimes held to point to a primitive ancestry. It has been remarked that in Marsupials it is the fourth toe which is dominant in size, whereas in Ungulates it is the third. An attempt has been made to explain this on the view (reasonable enough in itself) of a tree-living ancestry for the group. A greater development of the fourth toe is, however, by no means a necessary character of arboreal creatures; the Primates themselves are an exception. Nor is this prevalence universal among the Marsupials;

in Myrmecobius (alone) is the third toe the longest; and no great difference can be detected between the third and fourth toes in the case of the genera Phascologale, Didelphys, and some others. Professor Leche compares the predominance of the fourth toe with the hyperphalangeal condition in the fourth toe of the embryo Crocodile, and considers it an archaic feature, not surpassed by the ancient characteristics of the Monotremata. Again it has been pointed out that in Phascologale and Perameles, the epistropheus (axis vertebra) has a separate rib as in Ornithorhynchus. In the third place, the likeness of the teeth of Myrmecobius to those of Ornithorhynchus is an argument in the same direction, which is furthermore supported by the great age (Mesozoic) of the Metatherian group, if we are right in regarding those extinct creatures as Marsupials.

We may now mention certain facts which are not so generally used. The partly primitive structure of the right auriculo-ventricular valve in the Monotremata has no counterpart in any Marsupial which has been dissected; but there are traces in the latter of the characteristic "ventral mesentery" of Ornithorhynchus and Echidna.[^[65]] Mr. Caldwell's interesting observation upon the segmenting egg of the Marsupial, the incompleteness of the first segmentation furrow (reminding us of the meroblastic ovum of the Monotreme), may possibly not turn out to be so exclusively Marsupial a feature as has been thought.

The balance of evidence thus points to the nearer relationship of the Marsupials to the Eutherian mammals; and their great specialisation combined with certain evidences of degeneration (disappearance in part of the milk dentition), and their age, point to the fact that they are, at any rate, the descendants of an early form of Eutherian. But they must have separated from the Eutherian stock after it had acquired a definite diphyodonty and the allantoic placenta, the two principal features of the Eutherian as opposed to the Prototherian mammals.

Nevertheless it seems probable that the Marsupial tribe is derived from some of the earliest Eutherians. And on this view may be explained the retention of Prototherian characters.

The remaining Eutheria are obviously all to be referred to one great division with the possible exception of the Whales, whose affinities form one of the principal difficulties to the student

of this group. A short résumé of what is at present thought of the systematic position of this anomalous order is appropriate here. Albrecht went so far as to regard the Cetacea as the nearest group of animals to the hypothetical Promammalia.[^[66]] But discounting his arguments by the removal of such of them as relate to structure plainly altered by the singular mode of life of these creatures, there is really a great deal to be said in favour of his view.

The chief facts which argue a primitive position among mammals for the Cetacea are perhaps: (1) the slight union of the rami of the lower jaw; (2) the occasionally rather marked traces of the double constitution of the sternum; (3) the long and simple lungs; (4) the retention of the testes within the body-cavity; (5) the occasional presence (in Balaenoptera) of a separate supra-angular bone. These points, however, are but few, and are not of such great weight as those which ought to be present to establish a claim to separate treatment for the Cetacea as opposed to the Eutheria. If this group of mammals can be tacked on anywhere, it appears to us that the nearest relatives are not, as is sometimes put forward, the Ungulata or the Carnivora, but the Edentata. There are quite a number of rather striking features in which a likeness is shown between these apparently diverse orders of mammals. The chief ones are these: (1) the existence of traces of a hard exoskeleton, of which vestiges remain in the Porpoise; (2) the double articulation of the rib of the Balaenopterids to the sternum, with which compare the conditions obtaining in the Great Anteater; (3) the concrescence of some of the cervical vertebrae; (4) the share which the pterygoids may take in the formation of the hard palate; (5) the fact that in the Porpoise, at any rate, as in many Edentates, the vena cava, instead of increasing in size as it approaches the liver, diminishes.

Another group which is perfectly isolated is that of the Sirenia. The alliance advocated by some with the Cetacea, and quite recently renewed by Professor Haeckel, is contradicted by so many important features that it seems necessary to abandon it. The recent discovery of a fossil Sirenian jaw by Dr. Lydekker with teeth highly suggestive of those of Artiodactyla, may prove a clue. A third group which is so isolated as to have been placed in a

primary division, proposed to be called Paratheria, is that of the Edentates. Probably the group so called should really be divided into the Edentata and the Effodientia, the latter containing the Old World forms. Whether or not it be ultimately shown that the Ganodonta are ancestral Edentates (sensu strictiori), the connexion of the group with others is not at present plain. The same is the case with the extensive order of Rodents. It is true that the extinct order of the Tillodontia shows certain Rodent-like characters on the one hand, and likenesses to Ungulates on the other. Certain likenesses shown by such apparently diverse animals as the Rabbit and the Elephant used to be insisted upon by Professor Huxley. For the present, however, the Rodents must remain as an isolated group with only very dubious affinities to others. The remaining groups of existing mammals are easier to connect. At first the differences between a Cat and a Horse seem to be quite as wide as those which separate any two of the higher Eutherian orders. But it seems to become clearer and clearer, as palaeontological investigation proceeds, that the bulk of the Ungulate and the Carnivorous, Insectivorous, and perhaps Lemuroid stocks converge into the early Eocene Creodonta. From the Lemuroid branch the higher Primates can be derived. The only "Ungulates" which cannot be fitted in with some reasonable probability is the group of the Proboscidea. But of the early forms of this division we have at present no knowledge.

CHAPTER VII

EUTHERIA—MARSUPIALIA

Order I. MARSUPIALIA[^[67]]

The Marsupials may be thus defined:—Terrestrial, arboreal, or burrowing (rarely aquatic) mammals, with furry integuments; palate generally somewhat imperfectly ossified; jugal bone reaching as far as the glenoid cavity; angle of lower jaw nearly always inflected. The clavicle is developed. Arising from the pubes are well developed and ossified epipubic bones. Fourth toe usually the most pronounced. Teeth often exceed the typical Eutherian number of forty-four; molars generally four on each side of each jaw. As a rule but one tooth of the milk set is functional, which is (according to many) the fourth premolar. Teats lying within a pouch, in which the young are placed. Young born in an imperfect condition, and showing certain larval characters. There is a shallow cloaca. The testes are extra-abdominal, but hang in front of the penis. In the brain the cerebellum is completely exposed; the hemispheres are furrowed, but the corpus callosum is rudimentary. An allantoic placenta is rarely present.

Structurally the Marsupials are somewhat intermediate between the Prototheria and the more typical Eutheria, with a greater resemblance to the latter.

Fig. 59.—Rock Wallaby (Petrogale xanthopus), with young in pouch. × 1⁄7. (After Vogt and Specht.)

The name Marsupial indicates what is perhaps the most salient character of this order. The pouch in which the young are carried is almost universally present. It is less developed

on the whole in the Polyprotodont forms, such as the Thylacine, Dasyures, etc., but is found in so many of them that the two divisions of the Marsupials, the Diprotodonts and the Polyprotodonts, cannot be raised to distinct orders on this and other grounds. The marsupial pouch of the Marsupials must not, as has been already pointed out, be confounded with the pouch of the Monotreme mammals. Distinct teats are found in the marsupium of the Marsupials, while there are none in the mammary pouch of the Monotreme, the pouch itself indeed representing an undifferentiated teat, of which the walls have not closed up. The pouch opens forward in the Kangaroos, and backwards in the Phalangers and in the Polyprotodonts. Its walls are supported by a pair of bones diverging from each other in a

-shaped manner; these are cartilaginous and vestigial in the Thylacine. They

are the precise equivalents of similar bones in the Monotremata. It has been held, but apparently erroneously, that these bones are mere ossifications in the tendons of the external oblique muscle of the abdomen, or of the pyramidalis of the same region; and vestiges have been asserted to exist in the Dog. Such bonelets are undoubtedly present in the Dog; but it seems clear from their development in Marsupials, as structures actually continuous with the median unossified portion of the symphysis pubis, that the "marsupial bones" belong to that part of the skeleton, and that they correspond with the epipubis of certain amphibians and reptiles. The pouch, it may be remarked, exists in a rudimentary form in the males of many Marsupials.

Fig. 60.—Ventral surface of innominate bone of Kangaroo (Macropus major). × ⅓. a, Acetabulum; ab, acetabular border of ilium; is, iliac surface; m, "marsupial" bone; pb, pubic border; pt, pectineal tubercle; s, symphysis; si, supra-iliac border; ss, sacral surface; thf, thyroid foramen; ti, tuberosity of ischium. (From Flower's Osteology.)

Fig. 61.—Mammary foetus of Kangaroo attached to the teat. (Nat. size.) (From Parker and Haswell's Zoology.)

The most salient feature in the life-history of the Marsupials is the imperfect condition in which the young are born. The egg is no longer laid, as in the Monotremes; but curiously enough the ovum, which has the small size of that of the Eutheria, divides incompletely at the first division (as Mr. Caldwell has shown), and this developmental feature may perhaps be looked upon as a reminiscence of a former large-yolked condition. The young when born are small and nude; the newly born young of a large Kangaroo is perhaps as large as the little finger. The young are transferred by the lips of the mother to the pouch, where they are placed upon a teat. It is an interesting fact that they are not merely imperfect foetuses, but that they are actual larvae. They possess in fact at any rate one larval organ in the shape of

a special sucking mouth. This sucking mouth is an extra-uterine production, and is of course an adaptation to the particular needs of the young, just as are other larval organs, such as the chin-suckers of the tadpole, or the regular ciliated bands of the larvae of various marine invertebrate organisms.

There are a number of other features which distinguish the Marsupials from other mammals.

The cloaca of the Marsupials is somewhat reduced, but is still recognisable. Its margins in Tarsipes are even raised into a wall, which projects from the body.

The tooth series of the Marsupials was once held to consist of one dentition only, with the exception of the last premolar, which has a forerunner. The interpretation of the teeth of Marsupials are various. Perhaps most authorities regard the teeth as being of the milk dentition, with the exception of course of the single tooth that has an obvious forerunner. But there are some who hold that the teeth are of the permanent dentition. In any case it is proved that a set of rudimentary teeth are developed before those which persist. Those who believe in the persisting milk dentition describe these as prelacteal. Another matter of importance about the teeth of this order of mammals is that their numbers are sometimes in excess of the typical Eutherian 44. This, however, holds good of the Polyprotodonts only.

It was for a long time held that the Marsupials differed from all other mammals in having no allantoic placenta. But quite recently this supposed difference has been proved to be not universal by the discovery in Perameles of a true allantoic placenta. The Marsupials have been sometimes called the Didelphia. This is on account of the fact that the uterus and the vagina are double. Very frequently the two uteri fuse above, and from the point of junction an unpaired descending passage is formed (see Fig. 48 on p. [74]).

A character of the brain of Marsupials has been the subject of some controversy. Sir Richard Owen stated many years ago that they were to be distinguished from the higher mammals by the absence of the corpus callosum. Later still it was urged that a true corpus callosum, though a small one, was present; while, finally, Professor Symington[^[68]] seems to have shown that

the original statement of Owen was correct, at least in part. It is at most feebly developed (see Fig. 58, p. [118]).

As to skeletal characters, the Marsupial skull has on the whole a tendency towards a permanent separation of bones usually firmly ankylosed. Thus the orbitosphenoids remain distinct from the presphenoid. The palate is largely fenestrated, a return as it were—says Professor Parker—to the Schizognathous palate of the bird. The mandible is inflected; this familiar character of the Marsupials goes back to the earliest representatives of the order in Mesozoic times (see p. [96]); but it is not absolutely universal, being absent from the much weakened skull of Tarsipes. On the other hand, the inflection is nearly as great in certain Insectivores, in Otocyon, etc. The malar always extends back to form part of the glenoid cavity. The shoulder girdle has lost the large coracoid of Monotremes; this bone has the vestigial character that it possesses in other Eutheria. The clavicle is present except in the Peramelidae. A third trochanter upon the femur seems to be never present.

Fig. 62.—Skull of Rock Wallaby (Petrogale penicillata). (Ventral view.) ali, Alisphenoid; bas.oc, basi-occipital; bas.sph, basi-sphenoid; ex.oc, ex-occipital; ju, jugal; max, maxilla; pal, palatine; par.oc, paroccipital; p.max, premaxilla; pr.sph, presphenoid; pt, pterygoid; sq, squamosal; ty, tympanic. (From Parker and Haswell's Zoology.)

The Marsupials cannot be regarded as an intermediate stage in the origin of the Eutheria for a number of reasons. In the first place, the nature of their teeth shows them to be degenerate animals; one set, whether we regard it as the milk or permanent dentition, has become vestigial. The recent discovery of a true allantoic placenta in Perameles removes one reason for regarding

the Marsupials as primitive creatures. It implies on the whole that the Marsupials have sprung from a stock with an allantoic placenta. The alternative is to assume the independent development of an allantoic placenta in both groups of the Mammalia; unless indeed the genus Perameles is to be held to be the most primitive race of Marsupials living, a hypothesis which does not appear on the face of it likely. So long as it was believed that the mammary pouch of the Monotremes was the equivalent of the marsupium of the Marsupials, the persistence of this structure seemed to be a bond of union between the groups. But it is now known that the marsupium is a special organ confined to the Marsupials, an argument which is rather in favour of their being a lateral development of the mammalian stem. It is to be remarked also that the marsupium is feeblest in the Polyprotodonts, which may perhaps be looked upon as the most primitive of the Marsupials, owing to their more numerous teeth and other points to be referred to immediately.

Not only are the Marsupials interesting from the point of view of their structure; their present and past distribution is of equal interest. During the Mesozoic epoch they occurred in Europe and North America; but not, so far as negative evidence means anything, in Australia, which is now their headquarters. In Europe Marsupials lingered on into the Tertiary period, when they finally became extinct. In America, of course, the group has persisted to the present day. Now it is important to notice that the two main subdivisions of the Marsupials, the Polyprotodontia and the Diprotodontia, exist to-day in both Australia and South America. These two divisions, it should be explained, differ principally in that one has numerous, the other rarely more than two,[^[69]] incisors in the lower jaw. It is perhaps the more widely distributed opinion that the Polyprotodontia are the more archaic group; this opinion rests upon one or two facts in addition to the absence of specialisation in the incisor teeth. Among the Polyprotodontia the total number of teeth is greater—a clearly primitive character; secondly, the general form of the body of these animals, with four subequal limbs and carnivorous or omnivorous diet, contrasts with the purely vegetarian and much specialised Kangaroos at any rate. Finally—and sufficient stress

has perhaps not been laid upon this matter—the brain among the Polyprotodonts is less convoluted than among the genera of the other division. This statement is of course made with due regard to parallelism in size (see p. [77]). It is well known that the complexity of a brain bears a distinct relation to the size of its possessor within the group. Now the most ancient Marsupials are decidedly more Polyprotodont-like. No European form from the earlier periods is distinctly to be referred to the Diprotodonts. But both divisions now exist in America and Australia.

We must assume, therefore, one of three hypotheses. Either the differentiation into the two great divisions occurred in Jurassic or Cretaceous times before the migration of the order southwards; or the Diprotodont type is only a type, and not a natural group, i.e. it has been separately evolved in America and Australia; or, finally, there was formerly a land-connexion in the Antarctic hemisphere, along which the Diprotodonts of Australia wandered into South America. The middle hypothesis has this to commend it, that syndactylism occurs in both divisions, and that in some Diprotodonts the pouch opens backwards as it does in the Polyprotodonts. So great are the resemblances that but little difference is really left—of great importance that is to say. Hence it is not difficult to imagine the reduction of the incisors having taken place twice. In favour of the first hypothesis there are no positive facts. Finally, in favour of the last, which is so strongly supported by the facts of distribution derived from the study of other groups of animals,[^[70]] there is at least this striking fact or rather series of facts: that some of the South American fossil Polyprotodonts have a "strictly Dasyurine relationship."[^[71]] If there has not been a direct migration, then the Dasyurine type has been twice evolved, an improbability that few will attempt to explain away. In any case we shall adopt here the usual division of the Marsupials into Diprotodontia and Polyprotodontia.

Sub-Order 1. DIPROTODONTIA.

This group includes the herbivorous Marsupials. The incisors are as a rule three above, but one only in the Wombats. Below

is one strong pair, with occasionally one or two rudimentary incisors. The upper canines, if present, are not large. The molars are tuberculate or ridged. All Marsupials (except the Wombats) to some extent, and the Macropods especially, are characterised by the prolongation of the tubes of the dentine into the clear enamel. The significance of this fact is, however, lessened by the fact that the same penetration of the enamel by dentinal tubes occurs in the Jerboa, the Hyrax, and some Shrews. The feet have two syndactylous toes,[^[72]] less marked in the Wombats than in the Kangaroos and Phalangers.

Fig. 63.—Skull of Wombat (Phascolomys wombat). (Lateral view.) ang, Angular process; cond, condyle of mandible; ext.aud, opening of bony auditory meatus; ex.oc, exoccipital; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; p.max, premaxilla; sq, squamosal; ty, tympanic. (From Parker and Haswell's Zoology.)

This order is mainly Australian at the present day, using the term of course in the "regional" sense (see p. [84]); the only exception indeed to this statement is the occurrence of the genus Caenolestes in South America. But it is now known that Diprotodont Marsupials formerly existed in the same part of the world.

Fig. 64.—Bones of right foot of Kangaroo (Macropus bennetti). a, Astragalus; c, calcaneum; cb, cuboid; e³, ento-cuneiform; n, navicular; II-V, second to fifth toes. (From Flower's Osteology.)

Fig. 65.—Skeleton of Wallaby (Macropus ualabatus). The scapula is raised somewhat higher than in nature. The end of the tail is omitted. The head of the femur has been separated from the acetabulum. acet, Acetabulum; acr, acromion process; ast, astragalus; cal, calcaneum; cbd, cuboid; chev, chevron-bones; cl, clavicle; cun, cuneiform of carpus; epi, epipubis; fb, fibula; fem, femur; hd, head of femur; hu, humerus; il, ilium; isch, ischium; obt, obturator-foramen; orb, orbit; pis, pisiform; pub, pubis; rad, radius; rb¹, first rib; rb¹³, last rib; sc, scapula; st, sternum; tb, tibia; troch, great trochanter of femur; uln, ulna; unc, unciform; IV, fourth toe. (From Parker and Haswell's Zoology.)

Fam. 1. Macropodidae.—This family contains the Kangaroos, Wallabies, Rat-Kangaroos, and Tree-Kangaroos. With the exception of Dendrolagus the family is terrestrial, and its numerous species progress by leaps effected by the long hind-limbs, which are decidedly, often greatly, longer than the fore-limbs. In the hind-limb the fourth toe is very long and strong; the fifth moderately so; the second and third are slender and united by skin. The tail is always long, but differs in its characters from

genus to genus. The stomach is much sacculated. The dental formula is I 3/1 C (1 or 0)/0 P 2/2 M 4/4. The atlas is often open below, forming thus an incomplete ring.

Though the number of the incisor teeth in the adult Diprotodonts is never more than three on each side in each jaw, more numerous rudiments are present. Mr. M. Woodward[^[73]] has lately investigated the subject with interesting results. He finds that many species present decided traces of two additional incisors, raising the total to that which characterises the Polyprotodontia; but in two cases, viz. Macropus giganteus and Petrogale penicillata, a sixth is present, the total number being thus in excess of that found in any other Marsupial. This, as the author himself admits, proves too much. No mammal is known which in the adult condition has so many incisors; nor do the fossil Mammalia help us to get over the difficulty; even among reptiles it is not usual for so many teeth to occur upon the premaxillaries.

It is a curious fact that the two long lower incisors can be used after the fashion of a pair of scissors, or rather a pair of shears. Their inner edges are sharpened, and they are capable of some motion towards and away from each other; by their means grass is cropped.

The stomach of Macropus (and of other allied genera) is peculiar by reason of its long and sacculated character; the oesophagus enters it very near the cardiac end, which is bifid. Messrs. Schäfer and Williams[^[74]] have shown that the squamous, non-glandular epithelium of the oesophagus extends over the greater part of the stomach, only the pyloric extremity and one of the two cardiac caeca being lined with columnar epithelium.

The Macropodidae are clearly divisible into three sub-families, which are distinguished by marked anatomical characters.

In the sub-family Macropodinae (including the genera Macropus, Petrogale, Lagorchestes, Dorcopsis, Dendrolagus, Onychogale, and Lagostrophus) there is no hallux, and the tail is hairy. The oesophagus enters the stomach near the cardiac end. The caecum when short has no longitudinal bands; the liver has a Spigelian lobe.

The second sub-family, Potoroinae or Hypsiprymninae (including the genera Potorous, Aepyprymnus, Bettongia, and

Caloprymnus), consists of smaller animals than the Macropodinae, which, however, resemble them in having no hallux, but a hairy tail. The oesophagus enters the stomach near the pyloric end of that organ. The caecum, though short, has lateral longitudinal bands. The liver has no special Spigelian lobe. The canines are always present, being rarely so in Macropodinae, and are usually well developed.

The third sub-family, that of the Hypsiprymnodontidae, is doubtfully referable to the family; it consists of but one genus Hypsiprymnodon, which is in many points more like a Phalanger than a Kangaroo. It has an opposable hallux and a non-hairy, but scaly, tail. It has canines in the upper jaw.

Fig. 66.—Red Kangaroo. Macropus rufus. × 1⁄18.

Sub-Fam. 1. Macropodinae.—The genus Macropus includes not only the Kangaroos but also the Wallabies, which are really indistinguishable, though they have sometimes been placed in a separate genus Halmaturus. The genus thus enlarged contains twenty-three species. It may be thus characterised: the ears are long, the rhinarium is usually naked, but in M. giganteus and others a band of hairs descends to the upper lip; a naked band extends from the ankle to the pads on the digits, which is interrupted in M. rufus by a band of hairs just in front of the digits. The mammae are four. The tail is not bushy,

but is crested in M. irma. They are for the most part found on the Australian continent, but some species are found in the islands to the north which belong to the Australian region. Thus M. brunii, which is of interest as the first Kangaroo seen by a European, is a native of the Aru islands. A specimen of this animal, which was then living in the garden of the Dutch governor of Batavia, was described by Bruyn in the year 1711. M. rufus, the largest member of the group, is remarkable for the red secretion which adorns the neck of the male. It is caused by particles which have the appearance and colour of carmine. M. giganteus is not, as its specific name might imply, the "giant" of the race; its dimensions are given as 5 feet, while M. rufus is said to attain a length of 5 feet 5 inches, exclusive (in both cases) of the tail.

The account which Sir Joseph Banks gives[^[75]] in his diary of the Kangaroo is interesting, since he was one of the first naturalists to see that creature. In July 1770 it was reported to him that an "animal as large as a greyhound, of a mouse colour, and very swift" had been seen by his people. A little later he was surprised to observe that the animal "went only upon two legs, making vast bounds just as the jerboa does." The second lieutenant killed one of these Kangaroos, of which Sir Joseph Banks wrote that "to compare it to any European animal would be impossible, as it has not the least resemblance to any one I have seen. Its fore-limbs are extremely short and of no use to it in walking; its hind, again, as disproportionately long; with these it hops seven or eight feet at a time, in the same manner as the jerboa, to which animal indeed it bears much resemblance, except in size, this being in weight 38 lbs., and the jerboa no larger than a common rat." The beast was killed and eaten, and proved excellent meat. Sir Joseph Banks' observations upon the leaping of the Kangaroo are of interest, because it is often asserted that the tail is largely made use of as a third foot or as a support. Mr. Aflalo declares in the most positive way that after repeatedly examining the tracks upon soft sand immediately after the animal had passed, not the very faintest trace of the impression of the tail could be discovered. The leaps of a big Kangaroo seem to be somewhat greater than is recorded

by Banks. It is said that 15 or even 20 feet are covered at a bound, and in bound after bound. But in walking slowly it can be readily seen from an inspection of Kangaroos at the Zoological Society's Gardens that the animal does rest upon its tail, which with the hind-legs forms a tripod.

Petrogale with six species comes next to Macropus, and is indeed only to be differentiated from it by the thickly-haired and more slender tail, which is not used, as it is sometimes in the Kangaroos, as an extra hind-limb. The Rock-Kangaroos live among rocks, which they climb, and from which they leap; and the tail acts rather as a balancing pole. The most elaborate account of the anatomy of Petrogale known to me is by Mr. Parsons.[^[76]] The dentition as given by Mr. Thomas is I 3/1 C 0/0 Pm 2/2 M 4/4—that of Macropus without the occasionally occurring canine of the upper jaw. The osteological characters which separate it from Macropus are quite insignificant. Mr. Parsons mentions a wormian bone, "os epilepticum," at the junction of the coronal and sagittal sutures. It was found to occur in two out of five skulls examined, and appears not to occur in other Kangaroos. The palatine foramina of Petrogale are so large that the posterior part of the bone is only a narrow thickened ridge. The small intestine of P. xanthopus is 102 inches long, the large intestine 44 inches. The caecum has a length of 6 inches, and is not sacculated, differing in this from the caecum of Macropus major. The best known species are P. xanthopus and P. penicillata. The genus is confined to Australia itself, and does not enter Tasmania.

Onychogale includes the so-called "Nail-tailed Wallabies," which have a thorn at the end of the tail, reminding one of the Lion and the Leopard, whose tails have a similar armature. The muffle is hairy. Three species are allowed by Mr. Thomas.

Lagorchestes has, like the last genus, the rhinarium, i.e. that part of the nose immediately surrounding the nostrils, hairy instead of smooth as in the Kangaroos proper. It is distinguished from Onychogale by the absence of the terminal callosity to the tail, which is rather short. The name Hare-Kangaroo is given to the members of this genus (three species) on account of their exceeding fleetness. This genus is limited to Australia itself. L. conspicillatus is said to present "a remarkable

resemblance to the English hare," and L. leporoides was so called by Gould on account of general appearance as well as face.

Dorcopsis has shorter hind-legs than Macropus, and a naked muffle. The ears are small. The structure of D. luctuosa has been studied by Garrod,[^[77]] who pointed out the existence of four enlarged hair follicles on the neck near the mandibular symphysis. These are, however, represented in the next genus Dendrolagus, and occur also in Petrogale. The limbs are not so disproportionate as in Macropus, and the tail is naked at the tip.

Dorcopsis and the next genus to be described, Dendrolagus, differ from Macropus and its immediate allies, Petrogale and Lagorchestes, in a number of anatomical points. In the first place, the premolars are twice the size of those of Macropus, and they have a characteristic pattern not observable in the Kangaroos. This consists of a median ridge (the whole tooth being rather prismatic in shape), with lateral ridges at right angles to it. The upper canines are developed, but are minute.

The stomach is not quite like that of Macropus, though built upon a similar plan. The blind cardiac extremity is a single, not a double cul-de-sac; in this it is like that of Petrogale. The distribution of the squamous, white, oesophageal epithelium is very much like that of Dendrolagus. In both genera the orifice of the oesophagus into the stomach is guarded by two strong longitudinal folds, which run for some distance towards the pylorus. In Dendrolagus, at any rate, this tract is bordered on each side by glandular patches. In Dendrolagus, moreover, the squamous epithelium does not extend into the cardiac cul-de-sac. This latter is separated from the rest of the stomach by two slightly diverging folds, which are faintly represented in Petrogale and in Halmaturus. In the last two genera the folds surrounding the oesophageal orifice are but slightly represented; better in Halmaturus than in Petrogale. But there are not the patches of glands already referred to. The small intestine of Dorcopsis is 97 inches in length, the large being 32, i.e. proportionately long, as in Marsupials generally. The small caecum (2½ inches) is not sacculated.

The spleen is Macropodine, being

-shaped or

-shaped. The differences between Dorcopsis and the evidently closely allied Dendrolagus will be further considered under the description of

the latter. Dorcopsis is confined to New Guinea, and contains three species, viz. D. muelleri, D. luctuosa, and D. macleani. D. muelleri has a striking resemblance to Macropus brunii, with which it has been confounded. Though intermediate between Macropus and Dendrolagus, these Kangaroos are not arboreal.

The genus Dendrolagus is remarkable for its un-kangaroo-like habit of living in trees. In accordance with this change of habit is a relative shortening of the hind-limbs, a feature which begins to be observable in Dorcopsis. "The general build," writes Mr. Thomas, "is of the ordinary mammalian proportions, not macropodiform at all." The muffle is not naked for the greater part, though the shortness of the hairs gives that effect. As in Dorcopsis, but not as in Macropus, the bulla tympani is not swollen. There are altogether five species, the fifth, D. bennetti, having been lately described from specimens living in the Zoological Society's Gardens.

Fig. 67.—Tree-Kangaroo. Dendrolagus bennetti. × 1⁄12.

The anatomy of this genus has been described by Owen for D. inustus,[^[78]] and by myself for D. bennetti. The stomach, which

has a single, not bifid, cul-de-sac, is sacculated by two principal bands and other subsidiary ones. Its internal structure has already been to some extent described. The spleen of D. bennetti is remarkable for the fact that it is not

-shaped, whereas D. inustus agrees with other Macropodines in the form of this organ. The small intestine of D. bennetti is 95 inches long, the large 38. The caecum appears to differ in the two species; it is smaller in D. bennetti, where it is only 2 inches in length. The most remarkable feature of the liver is the large size of the left lateral lobe and the bilobed condition of the Spigelian lobe; this at least was the case with D. bennetti. A recently-described species[^[79]] has been attentively studied in its native haunts by Dr. Lumholtz.[^[80]] It lives in the highest parts of the mountainous scrubs of Queensland, where it moves quickly on the ground as well as among the trees. It is hunted with Dingos by the "blacks," and is eaten by them.[^[81]]

Lagostrophus is a generic name that has been proposed by Mr. Thomas for a small Wallaby 18 inches in length, which is distinguished by the fact that the long claws of the hind-limbs are entirely hidden by long and bristly hairs; the muffle is naked; there is no canine. The bullae are swollen. There is but one species of the genus, L. fasciatus, a native of West Australia.

Sub-Fam. 2. Potoroinae.—Aepyprymnus and the other genera placed in this sub-family are known by the vernacular name of Rat-Kangaroos, or sometimes Kangaroo-Rats. The latter term has been called "incorrect," though it is just as good as the former, both of them in fact being inaccurate as implying some likeness to or relation with a Rat. The present genus has a partially hairy rhinarium; the auditory bullae are not swollen. It contains but one species, Ae. rufescens, a native of Eastern Australia, which is distinguished by its very long hind-feet.

Bettongia has long hind-feet as in Aepyprymnus, but the rhinarium is entirely naked instead of being partially hairy, while the ears are much shorter. The genus, which contains four species, is remarkable as being the only ground-living mammal with a prehensile tail, which it uses to carry grass, etc.

B. lesueuri burrows in the ground, often to so great a depth as 10 feet. The genus occurs in Tasmania as well as in Australia.

Caloprymnus, with one species, is a genus instituted by Mr. Thomas in his Catalogue of Marsupials for a form (C. campestris) which combines in a remarkable way the characters of Aepyprymnus, Bettongia, and Potorous. The external characters and the general shape of the skull are as in Bettongia, while the molars have the structure of those of Aepyprymnus. The last premolar is as in Potorous.

Of the genus Potorous there are three species, which are Tasmanian as well as Australian. Unlike the other Rat-Kangaroos, the hind-feet are comparatively short, and the animal is therefore less addicted to jumping than its relatives. The rhinarium is naked, and the ears are of fair length.

Sub-Fam. 3. Hypsiprymnodontinae.—The Musk-Kangaroo, Hypsiprymnodon, is the last genus of the present family, and the only genus of this sub-family. It is intermediate between the Macropodidae and the Phalangeridae, the annectant character being mainly the hind-feet, which though they have the same long fourth digit as the Kangaroos, have it more feebly developed, and possess also an opposable hallux, which is one of the salient features in the structure of the Phalangeridae. The tail is naked and scaly; the rhinarium is entirely naked. The ears are large and not furry. The single species, H. moschatus, appears to feed upon insects as well as vegetables.

"Its habits are chiefly diurnal, and its actions when not disturbed by no means ungraceful. It progresses in much the same manner as the Kangaroo-Rats (Potorous), to which it is closely allied, but procures its food by turning over the débris in the scrubs in search of insects, worms, and tuberous roots, frequently eating the palm berries, which it holds in its fore-paws after the manner of the Phalangers, sitting up on its haunches, or sometimes digging like the bandicoots." This is Mr. Ramsay's description of the animal, which he was the first to discover.[^[82]]

Fam. 2. Phalangeridae.—The genus Hypsiprymnodon bridges over the not very wide gap which separates the Kangaroos from the Phalangers. The Phalangers are Marsupials with five fingers and toes; the second and third toes are bound together by a

common integument as in the Macropodidae. The hallux is opposable and nailless. The tail is nearly always long and prehensile. The pouch is well developed; the stomach not sacculated; a caecum is present (except in Tarsipes). These are really the principal distinctions between the two families. In addition, it may be mentioned that the lower incisors have not a scissor-like action as in the Kangaroos.

The Phalangers may be divided into four sub-families.

The first of these, that of the Phalangerinae, contains the genera Phalanger (including Cuscus), Acrobates, Distaechurus, Dromicia, Gymnobelideus, Petaurus, Petauroides, Dactylopsila, Pseudochirus, and Trichosurus.

These genera agree in the following generalities:—Tail well developed, often very long; three incisors above, and at least two premolars both above and below; caecum long and simple; stomach without a cardiac gland; liver not very complicated by secondary furrows, with a distinct caudate lobe; the vaginal median culs-de-sac often coalesced; lungs with an azygos lobe.

Fig. 68.—Bones of leg and foot of Phalanger. ast, Astragalus; calc, calcaneum; cub, cuboid; ect.cun, ecto-cuneiform; ent.cun, ento-cuneiform; fb, fibula; mes.cun, meso-cuneiform; nav, navicular; tib, tibia; I-V, first to fifth toes. (After Owen.)

The second sub-family, Phascolarctinae (with the Koala only), is thus characterised:—Tail rudimentary; cheek-pouches present; superior incisors three, but only one premolar above and below;

caecum extraordinarily long; stomach with a cardiac gland; liver complicated by additional furrows, without a free caudate lobe; no azygos lobe to lungs; vaginal culs-de-sac free.

The third sub-family, Phascolomyinae, contrasts with the others as follows:—Tail rudimentary; cheek-pouches present, but rudimentary; one incisor on each side above, but no additional premolars; all the teeth rootless; caecum not peculiar in shape; stomach with a cardiac gland; liver complicated by secondary furrows, without a free caudate lobe; lung with an azygos lobe; vaginal culs-de-sac free.

The last sub-family, Tarsipedinae, is thus defined:—Tail long; tongue extensile; only one premolar; molars reduced; caecum absent.

Fig. 69.—Vulpine Phalanger. Trichosurus vulpecula. × 1⁄6.

Sub-Fam. 1. Phalangerinae.—The genus Phalanger embraces five species, sometimes called by the generic name of Cuscus. They are largish animals with short ears; only the end of the tail is naked. Of these animals only one species is found in Australia itself, the rest inhabiting the islands lying to the north. The Spotted Cuscus, Ph. maculatus, is in spite of its vegetarian diet, and perhaps on account of its spots, spoken of as the "Tiger Cat." Mr. Aflalo remarks of it that though provided with a prehensile tail, it is little better as a climber than the tailless Koala.

Trichosurus, including the "True Phalangers," includes largish species, which can be distinguished from the last genus by a chest-gland similar to that which occurs in Myrmecobius and some other Marsupials of the present group. There are but two species, which are purely Australian. The "Brush-tailed Opossum," T. vulpecula (perhaps better known as Phalangista

vulpina), like its American pseudo-namesake (a true Opossum, genus Didelphys), "plays 'possum" on occasions. The dental formula is I 3/2 C 1/0 Pm 2/3 M 4/4. The ears are shortish.

The Ring-tailed Phalangers, Pseudochirus, are more widely distributed than the last two genera; they range from Tasmania in the south to New Guinea in the north. They are not, however, ring-tailed, though the tip of the tail is generally white. As in the last genera, which have prehensile tails, the end of this appendage is naked. The mammae are four. The tooth formula is I 3/2 C 1/0 Pm 3/3 M 4/4. There are some ten species of the genus.

The Striped Phalanger, Dactylopsila trivirgata, is an animal about a foot long, whose identity can be ascertained by its striped, black and white skin. It is an arboreal creature that lives apparently both on leaves and grubs like so many arboreal creatures of quite different groups—Squirrels, for instance, and New-World Monkeys. The tooth formula is I 3/3 C 1/6 Pm 3/2 M 4/4.

Gymnobelideus leadbeateri is a small creature with a body 6 inches in length. It is restricted to the colony of Victoria. The general look is that of Petaurus; the ears are naked.

Dromicia is a genus of Phalangers which although devoid of a parachute, such as is possessed by certain genera that will be considered immediately, is able to leap with great agility from branch to branch. The ears are large and thin and almost naked; the tooth formula is I 3/2 C 1/0 Pm 3/3 M 4/4. They are minute creatures, the longest measuring, with the tail, but 10 inches. Dormouse-Phalanger is a name sometimes given to them. There are four species, ranging from Tasmania to New Guinea. The name Dormouse as applied to the genus seems to be owing to the way in which they hold a nut in the paws when feeding. D. nana is 4 inches long, with a tail of nearly the same length. It is thick at the base.

Distaechurus is the last genus of non-flying Phalangers. Its name refers to the arrangement of the hairs on the tail, which are disposed on either side in a row like the vane of a feather. The tooth formula is I 3/2 C 1/0 Pm 3/2 M 3/3, very nearly as in Acrobates. The ears are as in that genus.

Petaurus is the first genus of the Flying Phalangers, all of which are provided with a parachute-like expansion of the skin between the fore- and hind-limbs; the ears are large and naked; and the tooth formula is I 3/2 C 1/0 Pm 3/3 M 4/4. There are three

species of the genus, which extend through pretty well the entire Australian region. The term "flying" as applied to these and the other "flying" genera is of course an exaggeration. The animals cannot fly upwards; they can only descend in a skimming fashion, the folds of skin breaking their fall. P. breviceps is perhaps the best-known species. The body is 8, the tail 9 inches long.

Petauroides seems to be chiefly distinguished from Petaurus by the fact that, as in its ally Dactylopsila, the tail is partly naked terminally. In Petaurus and Gymnobelideus the tail is bushy to the very end, including its extreme tip below.

A third genus of Flying Phalangers is the minute Acrobates, which has a distichous tail like that of Distaechurus. It is not more than 6 inches in length including the tail. As to these Flying Phalangers it is exceedingly instructive to observe that the same method of "flight" has been apparently evolved three times; for the three genera are each of them specially related to a separate type of non-flying Phalanger. The same observation can be made about the Flying Squirrels, Anomalurus and Sciuropterus. The dental formula is I 3/2 C 1/0 Pm 3/3 M 3/3. The ears are thinly clad with hair. There are four teats.

Sub-Fam. 2. Phascolarctinae.—The Koala, or Native Bear, Phascolarctos cinereus, is the only representative of its sub-family. It is, like the Wombat, aberrant in the lack of an obvious tail. The absence of this appendage is curious in an arboreal creature whose near allies have a long and prehensile one. The structure of the Koala was investigated by the late Mr. W. A. Forbes.[^[83]] There are some unexpected points of likeness to the Wombat: thus they agree in the absence of the tail, in the structure of the stomach, and in the great subdivision of the lobes of the liver. The brain, however, is smooth, and the caecum is exceedingly large and complicated in structure, that of the Wombat being short. That both animals have cheek-pouches is perhaps due to similar habits of temporarily storing masses of food. This animal has only eleven pairs of ribs. The tail has only seven or eight vertebra, and these have no chevron-bones.

A peculiarity of the skull is seen in the great size of the alisphenoid bulla, which is comparable in size and appearance with that of the Pig. As in the Kangaroos, the atlas is incomplete below.

The tooth formula of the genus is I 3/1 C 1/0 Pm 1/1 M 4/(4 or 5). The additional lower molar seems to be exceptional, and has been found in one specimen only.

In the alimentary tract the most remarkable structure is the large intestine, which is very capacious for the first 28 inches or so of its course. This section of the colon is lined with rugae precisely like those which are found in the caecum. These folds, which at first are some twelve in number, fuse lower down, and by the time that the colon approaches the external orifice are reduced to five. Similar folds, as already stated, occur in the caecum, but do not extend as far as its blind end. The caecum is proportionately and actually larger than in any other Marsupial. The gall-bladder is unusually elongated.

Fig. 70.—Koala. Phascolarctos cinereus. × 1⁄9.

The Koala is mainly crepuscular or nocturnal in its habits. It feeds so exclusively upon the leaves of the gum-tree (Eucalyptus) that it is impossible to keep the creature long in captivity in lands where that particular kind of food is not available.

The female, though she seems to bear but a single young one, which is carried on the back after the fashion of some Opossums, has two nipples. The animal's slow habits seem to require a nocturnal and retired life. It is about as lethargic as the Sloth, and it is said to further resemble that animal in clinging firmly to a branch even after it is shot.

Fig. 71.—Wombat. Phascolomys wombat. × 1⁄12.

Sub-Fam. 3. Phascolomyinae.—Phascolomys, the Wombat, is the only genus of this sub-family. This animal has the appearance of a heavily-built Marmot, like which it has a mere stump for a tail, and a pair of strong chisel-shaped and Rodent-like incisors, which, however, differ from those of Rodents in having a complete coating of cement. All the teeth of the animal are rootless, and there are no canines. The incisors have enamel on the front and lateral faces only. The dental formula is I 1/1 C 0/0 Pm 1/1 M 4/4. The affinities with other Diprotodont Marsupials are shown by the commencing syndactyly of the second and third toes. The rhinarium is naked or hairy. There is a rudimentary cheek-pouch, as in Phascolarctos. The Wombat has, like the Koala, and also the Beaver—which does away with some of the value of the comparison—a peculiar gland-patch in the stomach, a raised area of collected glands. In no other Marsupial is such a structure found, "whilst in the two forms under consideration its identity is almost precise. That such a unique structure should have been independently developed in two forms unrelated to each other, appears to me to be in the highest degree improbable." This is Mr. Forbes' opinion. It might be strengthened by adding the observation that, as there are other points of likeness between the Wombat and the Koala, it seems more unlikely that a structure so nearly identical should have been twice

developed in two not very distant forms. As in the Kangaroos, the atlas is open below. Ph. ursinus has 15 ribs; the other species the normal (for Marsupials) 13. Other points of likeness will be mentioned under the description of the Koala. These animals mainly feed upon roots; they live in companies in burrows. There are three species—Ph. ursinus, Ph. latifrons, and Ph. mitchelli. Ph. ursinus is Tasmanian in range, the other two species South Australian.

Fig. 72.—Skull of Wombat. Phascolomys wombat. (Lateral view.) ang, Angular process; cond, condyle of mandible; ex.oc, exoccipital; ext.aud, opening of bony auditory meatus; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; p.max, premaxilla; sq, squamosal; ty, tympanic. (From Parker and Haswell's Zoology.)

Sub-Fam. 4. Tarsipedinae.—The genus Tarsipes ought perhaps to be removed from the present family. There is but a single species, which is a small creature of 7 inches in total length, of which the tail measures 4 inches. The teeth are much dwindled, the formula being I 2/1 C 1/0 Pm 1/0 M 3/3 = 22. The lower incisors are procumbent. The lower jaw, moreover, has not the characteristic Marsupial inflection. The intestinal canal is without the caecum present in the remaining Phalangeridae. It is a curious fact that this aberrant little Phalanger should come from Western Australia, like the even more aberrant Myrmecobius. Like the latter also, Tarsipes has a long exsertile tongue, with which, however, it extracts honey from flowers. Probably it also catches minute insects in the corollas of the flowers. It has been proved, in fact, that in captivity at any rate the animal is insectivorous; for it has been known to eat moths.

Fam. 3. Epanorthidae.—The extinct Epanorthidae of

Patagonia are represented to-day by a small Marsupial which has been rediscovered within the last two or three years. This little animal, formerly called Hyracodon (a pre-occupied name), is now termed Caenolestes, and is a native of Colombia and Ecuador. There are two species, and of these C. obscurus is called by the inhabitants "Raton runcho," which means opossum-rat. It lives apparently upon bird's eggs and small birds, though it belongs to the Diprotodont division of the Marsupials. Caenolestes, however, although diprotodont, has not the syndactylous character of the digits of the feet already referred to in the Kangaroos and their allies. The pouch is small and rudimentary. The dentition is I 4/3 C 1/1 Pm 3/3 M 4/4 = 46, and the teeth are said by Mr. Thomas to be much like those of the Australian Dromicia.[^[84]]

In the skull a peculiarity which does not bear upon its affinities to other Marsupials, but is still interesting, is mentioned by Mr. Thomas. The nasals are not sufficiently prolonged to meet the upper edge of the maxillae, and so a vacuity is left, as in the skulls of many Ruminants (e.g. the Sable Antelope). The palate is very imperfect; the foramina, which render it so, reach as far forward as the last premolar. The lower jaw has quite the appearance of that of a Macropus or Phalanger, with long and forwardly projecting incisors.

Extinct Diprotodonts.—The great Diprotodon is a creature with a skull a yard long, which must have been of the size of a large Rhinoceros. Though closely allied to Macropus, it seems that this great beast did not hop after the fashion of a Kangaroo, its limbs being of a more equal size than in the Kangaroo. Recently some further remains of Diprotodon have been discovered in a lake known as Lake Mulligan, where they had apparently been bogged. Professor Stirling has contributed an account of these remains, which fills up a considerable gap in our knowledge. He has been able to state the structure of the fore- and hind-limbs. Both limbs are pentadactyle, the fingers of the fore-limb being approximately equal in length and general development. In the hind-limb the hallux is small, and consists of the metatarsal only. This bone is fixed in the position of "extreme abduction," and is suggestive of an arboreal limb. Digits two and three may have

been syndactylous, and the authors of the account[^[85]] of these bones think that the fourth toe may have shared in this syndactyly. The metatarsal of the fifth digit is enormously expanded at its edge, and seems to have furnished a strong support to the creature; this is also seen in the metacarpal of the fore-limb. Probably, therefore, Diprotodon was quadrupedal in its mode of progression, with the emphasis laid upon the little finger and the little toe instead of, as in ourselves, the first toe. The hind-foot of the Diprotodon could not be more unlike that of a Kangaroo than it actually is.

Another giant among these Marsupials was the genus Thylacoleo, whose name was given to it by Sir Richard Owen on the view that it was a Marsupial Tiger. Sir W. Flower has, however, controverted this opinion, and the genus is in fact, in spite of its large size, closely allied to the Phalangers and

Cuscuses.[^[86]] The dental formula is I 3/1 C 1/0 Pm 3/1 M 1/2; the last premolar is a great blade-shaped tooth like that of Potorous.

Nototherium was a creature smaller than Diprotodon, but still of large size; it is believed to have been a burrowing creature, and to connect the Wombats with Diprotodon. More certainly allied to the existing Wombat was Phascolonus, a Wombat as big as a Tapir.

Fig. 75.—Nototherium mitchelli. Side view of skull. × 1⁄6. (After Owen.)

Of extinct American Diprotodonts the Epanorthidae, already referred to in connexion with the living Caenolestes, were the most prominent forms. The genus Epanorthus occurs in the Santa Cruz formation of Patagonia, which is believed to be Miocene. The incisors are three in the upper jaw; and the single incisor of each ramus of the lower jaw is a great chisel-shaped, cutting instrument.

Abderites is also typically Diprotodont by reason of the large projecting incisors of the lower jaw. It has a large cutting tooth in the lower jaw, which appears to be the last premolar, and is thus comparable to the great cutting tooth of the lower jaw and of the upper jaw of the extinct Phalanger, Thylacoleo.

It may also be comparable to the great premolar of such Multituberculata as Ptilodus and Plagiaulax. It is, moreover, marked with vertical grooves.

An interesting form, which is unfortunately but little known, is the Australian and Pleistocene genus Triclis, with one species, T. oscillans. In having a minute canine tooth in the lower jaw it agrees with some Phalangeridae, and being otherwise closely allied to Hypsiprymnodon, it unites the Macropodidae with the Phalangeridae.

Sub-Order 2. POLYPROTODONTIA.

In this mainly carnivorous or insectivorous division of the Marsupials the incisors are four or five on each side of the upper jaw, and one or two fewer in the lower jaw. Figs. 76 and 77 illustrate the Polyprotodont and Diprotodont dentitions. The canines are those of flesh-eaters and so are the molars, being as a rule sharply cuspidate. As a rule, which has an exception in the Peramelidae, there is no syndactylism of toes in the hind-foot. This sub-order is at the present day Australian and American in its range.

Fig. 76.—Front view of the skull of Tasmanian Devil (Sarcophilus ursinus), showing Polyprotodont and carnivorous dentition. (After Flower.)

Fam. 1. Dasyuridae.—This family consists of Marsupials which are generally pentadactylous, but with occasionally the hallux missing. The tail is long but not prehensile. The pouch is present or absent. The teeth vary in the different genera, but

the upper incisors are never less than three, and may be as many as five in the upper jaw and six in the lower. The canines are trenchant. There is no caecum.

Fig. 77.—Front view of skull of Koala (Phascolarctos cinereus), illustrating Diprotodont and herbivorous dentition. (From Flower.)

Fig. 78.—Longitudinal section of the skull of the Thylacine (Thylacinus cynocephalus). × ½. a, Angular process of mandible; AS, alisphenoid; BO, basioccipital; BS, basisphenoid; cd, condyle of mandible; ET, ethmoturbinal; Ex.O, exoccipital; Fr, frontal; ME, ossified portion of mesethmoid; MT, maxilloturbinal; Mx, maxilla; Na, nasal; OS, orbitosphenoid; Pa, parietal; Per, periotic; Pl, palatine; PMx, premaxilla; PS, presphenoid; Pt, pterygoid; SO, supraoccipital; Sq, squamosal; Vo, vomer. (From Flower's Osteology.)

The genus Thylacinus contains but a single species, which is now limited to Tasmania, and is generally known as the Tasmanian Wolf. It has the build of an ordinary Wolf, and is of about the same size. The hinder part of the body is marked with a series of black transverse bands. The hallux is entirely wanting; the pouch opens backwards. The marsupial bones are minute and unossified. The dental formula is I 4/3 C 1/1 Pm 3/3 M 4/4 = 46. There are four mammae. This animal, now confined to Tasmania,

is getting rarer on account of its sheep-killing propensities, and the consequent war of extermination declared upon it by the colonists. It will, however, feed upon other animals; and it is related that the first specimen ever captured had in its stomach the remains of an Echidna! Mr. Thomas thinks that the persistence of this and of some of the other larger carnivorous Marsupials in Tasmania after their extinction in Australia is not unconnected with the advent of the Dingo. But it is stated that the Thylacine is quite capable of keeping even a pack of dogs at bay.

Fig. 79.—Tasmanian Devil. Sarcophilus ursinus. × 1⁄10.

The genus Sarcophilus has been frequently confounded with the next, but it is kept apart by Mr. Thomas, who follows Cuvier in this. An alternative generic name is Diabolus, which, like the first name, refers to the habits and character of the single species which this genus contains. The genus is more like Thylacinus than is Dasyurus. The hallux is wanting, and the teeth, though fewer in number (42), resemble those of the Thylacine more closely than do those of the Dasyure. The species is called S. ursinus, the popular name being Tasmanian Devil. It is black with a variable number of white patches on the body. It is of about the size of a Badger, and is, like the Thylacine, a nocturnal animal. The Tasmanian Devil is said to be one of the most ferocious of animals, and to express its ferocity by a "yelling growl."

Fig. 80.—Skull of Dasyurus. (Lateral view.) al.sph, Alisphenoid; ang, angular process of mandible; fr, frontal; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; oc.cond, occipital condyle; par, parietal; par.oc, paroccipital process; p.max, premaxilla; s.oc, supraoccipital; sq, squamosal; sq′, zygomatic process of squamosal. (From Parker and Haswell's Zoology.)

Fig. 81.—Dasyure. Dasyurus viverrinus. × 1⁄5. (After Vogt and Specht.)

The next genus of this family, Dasyurus, comprises five species, which range over the whole of the Papuan and Australian sub-regions. The general form is Viverrine, and the hallux is sometimes present though small. The dental formula is as in the last genus, but the teeth "are more insectivorous in their character." There are six or eight mammae. The members of this genus are grey or brown, and spotted with white; they are all arboreal, and feed largely upon birds and their eggs. Mr. Thomas has pointed out that in two species, D. viverrinus and D. geoffroyi, the striae upon the foot-pads are absent, and that therefore these at least are probably not so purely arboreal as the rest. The animals are not diurnal, and during the day hide themselves in the hollow trunks of trees. They are spoken of as "Native Cats," but have the general habits of Martens. D. maculatus is common in Tasmania, but is rare in Australia, thus "approaching the condition now exhibited by the Thylacine and

Tasmanian Devil, namely, complete extermination in Australia, where both once lived." D. hallucatus shows an approach to Phascologale in its five-toed hind-feet and slender build.

Phascologale is a genus which, like the last, is usually arboreal (although not P. virginiae of North Queensland), but is of much smaller size, the species not exceeding the dimensions of a rat. They have no spots, but there is sometimes a stripe down the back. There are thirteen species, which have the same range as the last genus. The hallux is present though small, but the pouch is "practically obsolete," though there is a small fold of skin behind the teats. The rhinarium is naked; the tail is long, "bushy, crested, or nearly naked." The mammae are four to ten in number. The dental formula is as in Dasyurus, and the teeth are not very different in form; sometimes the last premolar is wanting. "The members of this genus," remarks Mr. Thomas, "evidently take the place in the Australian region filled in the Oriental by the Tupaiae, and in the Neotropical by the smaller Opossums."

The genus Sminthopsis comprises not more than four species, even smaller than the last. The largest species, S. virginiae, is only 125 mm. in length. The hallux is present, and there is a well-developed pouch. There are forty-six teeth, as in the Dasyures. The feet are narrow with granulated or hairy soles, whereas in Phascologale they are broad with smooth soles. The mammae are eight or ten. The genus ranges through Australia and Tasmania.

The genus Antechinomys has but a single species, which is a native of Queensland and New South Wales. The build is Jerboa-like, and the animal is, as might be inferred, terrestrial. The ears are very long, and the limbs elongated; the hallux is absent; the teeth are exactly as in Sminthopsis.

Antechinomys has thirteen dorsal and seven lumbar vertebrae; three sacrals and twenty-five caudals, the latter number being in excess of that of its allies. The stomach is nearly globular, with approximated orifices; the intestine measured 6.8 inches, a little more than twice the length of the animal itself. A. lanigera is a native of East Central Australia, and appears to be entirely terrestrial in habit, and to progress by a series of leaps—at any rate when going at full speed.

Professor Spencer, who found examples of this rare species, gives

an interesting description of its habits. Antechinomys has much the look of the Australian Rat, Hapalotis mitchelli; and as the two animals lead a similar kind of life, the resemblance is not unexpected. Professor Spencer wonders why these creatures are saltatory in habit. The country which they inhabit is arid, but with patches of grass and shrubs. For a big kangaroo the advantage of the power of leaping over such obstacles may be obvious, but not for the small and slender Antechinomys. The chief foes of this rare Marsupial appear to be predatory birds; and Professor Spencer thinks that the saltatory mode of progression may be more baffling to such pursuers than even a rapid run.

The genus Dasyuroides has been lately instituted by Professor Spencer for a Marsupial from Central Australia somewhat intermediate between Sminthopsis and Phascologale. As there is but one species, the generic will be considered with the specific characters. D. byrnei is an animal of about the size of the Common Rat. The hallux is absent. The tail is fairly thick, but not "incrassated." There are six mammae, and the pouch is but slightly developed, with two low lateral folds. The dentition is I 4/3 C 1/1 Pm 3/2 M 4/4. This Marsupial is nocturnal, and burrowing in habit. Its food consists of insects.[^[87]]

Myrmecobius is so different from the last-described genera (Dasyurinae) that it is usually separated from them as a sub-family Myrmecobiinae. The animal is of a bright rufous colour, banded posteriorly with white. There is no hallux, though the metatarsal belonging to that digit is present. There are four mammae.[^[88]] On the chest is a naked patch of some extent, upon which open the ducts of a complex gland, which has been described and figured by myself.[^[89]] There is no pouch, but a tract of skin shows indications of a pouch-like structure. The teeth are extraordinarily numerous, fifty to fifty-four; the formula being I 4/3(4) C 1/1 Pm 3/3 M 5/6. Their resemblance to those of certain Jurassic Marsupials is dealt with on p. [100].[^[90]] In this matter lies of

course the chief interest of the genus, which may be "an unmodified survivor from Mesozoic times, and therefore from a time long before the Didelphyidae, Peramelidae, and Dasyuridae were differentiated one from the other." Another ancient feature (found in Jurassic mammals) is a mylo-hyoid groove upon the lower jaw, which, however, is not always present, and its existence has therefore been denied. The single species, M. fasciatus, is partly arboreal and partly terrestrial in habit, and feeds upon ants. It is a Western and Southern Australian form.

Fig. 82.—Banded Australian Anteater. Myrmecobius fasciatus. × 1⁄5.

Fam. 2. Didelphyidae.—All the members of this family are pentadactylous. The teeth are fifty in number, arranged thus: I 5/4 C 1/1 Pm 3/3 M 4/4. The caecum is small; the pouch is generally absent; the tail generally long and prehensile.

Fig. 83.—Virginian Opossum. Didelphys virginiana. × 1⁄5. (After Vogt and Specht.)

The genus Didelphys contains most of the forms belonging to this family, including as it does some twenty-three species. The Opossums are mainly arboreal animals, insectivorous in their food; but the larger species eat reptiles, birds, and their eggs. Several of the small species carry their young, when able to leave the teats, on

their back, the tails of the young being wrapped round that of the mother. It is not only the pouched species which carry their young in something of this fashion. Azara's Opossum, an animal as big as a cat, is said to carry its eleven young ones (themselves as large as rats) on the back, though their foothold does not appear to be strengthened by intertwining the tails. Even with this huge family on her back, the mother can climb trees with considerable alacrity. The mammae are seven to twenty-five in number. The genus has been lately split up into a number of genera, Marmosa, Dromiciops, Peramys, etc.

Fig. 84.—Thick tailed Opossum. Didelphys crassicaudata. × 1⁄5.

Chironectes is hardly different from Didelphys. It has webbed hind-feet, and is aquatic in habit. The one species of the genus is known as the Yapock, and is a Central and South American form. It is of about the size of a large rat, and appears to be an expert diver after the fish upon which it lives.

Fam. 3. Peramelidae.—The Bandicoots, although clearly belonging to the Polyprotodont Marsupials, yet agree with the Diprotodonts in the fact that the second and third toes of the feet are bound up in a common integument, which is not the case with the Diprotodont Caenolestes. The hind-feet are longer than the front; of the former limb, two or three of the fingers alone are long and functional; the others are rudimentary or absent. Tail long, hairy, and non-prehensile. Dentition I 5/3 C 1/1 Pm 3/3 M 4/4 = 48, or sometimes, owing to the absence of a pair of upper incisors, 46. There is a caecum.

Fig. 85.—Bones of manus. A, of Choeropus castanotis. × 2. B, of Bandicoot (Perameles). × 1½. c, Cuneiform; l, lunar; m, magnum; R, radius; s, scaphoid; td, trapezoid; tm, trapezium; u, unciform; U, ulna; I-V, digits. (From Flower's Osteology.)

Fig. 86.—Rabbit Bandicoot. Peragale lagotis. × 1⁄5.

The genus Peragale, the Rabbit-Bandicoots, consists of two species entirely Australian in range. The enormous ears (whence "Rabbit" Bandicoot) distinguish this genus from Perameles. The pouch opens backwards, and there are eight mammae. P. lagotis, the only species about whose ways of life anything is

known, burrows in the soil, whence it extracts grubs; it is also a grass-feeder, and it is said that its likeness to a Rabbit in appearance is strengthened by its similarity in flavour!

Perameles is a genus consisting of twelve species, which are found in Tasmania, Australia, and New Guinea. Like the last genus, from which it does not widely differ in other points, Perameles consists of species which combine insectivorous and vegetarian habits. One species is said to become in captivity an expert in catching mice. The pouch opens backwards, and there are six or eight mammae.

Fig. 87.—Pig-footed Bandicoot. Choeropus castanotis. × ⅓.

The last genus of this family is Choeropus, containing but one species, Ch. castanotis. It is confined to the Australian continent. It is to be distinguished from the last two by the fact that there are only two functional digits, the second and third, in the fore-limb; the fourth is rudimentary; the other two are absent. It burrows, and is omnivorous like its allies. The two metacarpals that are developed are very long and closely apposed; they have hence a remarkably pig-like aspect, and justify its name. The pouch opens backwards, and there are eight mammae.

Fam. 4. Notoryctidae.—This family contains but a single genus and species, the recently-discovered Notoryctes typhlops.[^[91]]

We may regard as family-characters the pentadactyle limbs, the existence of three pairs of incisors in the lower and four in the upper jaw; and the tritubercular nature of the upper molars. Notoryctes typhlops, the "Marsupial Mole" as it has been termed, was originally discovered by Professor Stirling in Central South Australia. It is a burrowing creature, clothed in a silky fur of a pale golden red, without external ears. It has been compared in appearance with Chrysochloris, the Cape Golden Mole, and the eminent palaeontologist, Professor Cope, has even insisted upon a real genetic affinity. Edentate affinities have also been suggested. But Notoryctes has a small pouch opening backwards as in other Polyprotodonts,[^[92]] and as it also possesses marsupial bones it must undoubtedly be referred to the Marsupialia. The animal shows many curious adaptations to its underground mode of life. Certain of the vertebrae in the neck and in the lumbar region are firmly welded together, giving of course a strength of push, and suggesting the Armadillos; the claws of the third and fourth front-toes are greatly enlarged, and must be efficient digging organs. The track of the animal is like that of a railway in mountainous country; it burrows for a short distance, emerges, and then descending beneath the surface re-emerges. The red colour of the fur is said to be in harmony with the arid soil in which it lives. The native name of the creature is "Urquamata." It feeds upon ants and other insects.

Fig. 88.—Australian Marsupial Mole. Notoryctes typhlops. × ¼.

Extinct Polyprotodonts.—Of extinct Polyprotodonts (apart from those Mesozoic forms which are considered on p. [100]) extinct species of Thylacinus and Dasyurus are known from

Australia. The most interesting fact in connexion with the Tertiary Polyprotodonts is the existence in South America of such genera as Prothylacinus and Amphiproviverra, which are not merely Polyprotodonts but definitely Dasyures, and not referable to the Didelphyidae.

These forms have been included in an order, Sparassodonta. But it is not by any means certain whether these forms are rightly placed in the neighbourhood of the carnivorous Marsupials; it is possible that they ought to be relegated to the Creodonta or to their allies. Their structure is in fact somewhat intermediate between those two groups. The teeth seem to be carnivorous and Marsupial-like in form; but as already mentioned, in connexion with the general structure of teeth, more than a single premolar is replaced. These animals in fact, in so far as regards their teeth, are midway between the Marsupials and the typical Eutheria. The angle of the lower jaw is inflected, but the palate is not marked by deficient ossification. At least this is not the case with all the members of the group. Whether the small Microbiotherium, which is made the type of a family, is rightly referred here is not certain. This animal had palatine vacuities as well as an inflected angle to the lower jaw.

CHAPTER VIII

EDENTATA—GANODONTA

Order II. EDENTATA

Terrestrial, partly subterranean, or arboreal creatures of quite small to gigantic size (some extinct genera), with frequently a covering of scales or bony scutes. Limbs clawed. Teeth either totally absent or, if present, imperfect in structure, being without enamel, and not forming a complete series; incisors and canines being as a rule absent. Teats axillary, pectoral, or inguinal.[^[93]] Retia mirabilia very common in the extremities.

To this group the name of Bruta was given by Linnaeus, but then it included not only the families which we now place in the modern order Edentata, but also the Elephant and the genus Trichechus. Mr. Thomas has proposed to change the name into Paratheria, which name is suggestive of what he and some others think concerning the systematic position of the group, i.e. that it is not to be placed in the Eutherian group of mammals at all, but represents a separate twig which has arisen with the Eutheria from a low mammalian stock. This view can hardly be accepted if the Ganodonta—which will be treated of presently—be really ancestral Edentates, for they are not in any way a Prototherian mammalian group, so far as their remains enable us to judge.

The Edentata contain the Sloths, Ant-bears, Armadillos, Manis and Orycteropus, among living forms. The great Ground-Sloths, Megatherium, etc., and Armadillos, Glyptodon, etc., represent the extinct forms.

The name that has been applied to this group is inappropriate

inasmuch as many Edentates have teeth. It is, however, by a number of small tooth-characters that the order can be defined. Thus if teeth are present they are simple in structure, without enamel in the adult condition, though a rudimentary enamel-organ has been discovered in an Armadillo. The teeth, moreover, are not found in the anterior part of the mouth, and they grow from persistent pulps; neither is there much differentiation among them. It is not possible, however, to speak of the Edentates as quite homodont, since in Orycteropus there are large cheek-teeth; but there is at any rate not a marked heterodonty in that or in any other Edentate. It used to be said that the Edentates were monophyodont. But the Armadillo Tatusia was subsequently found to possess a second suppressed dentition, and after this discovery Mr. Thomas proved that Orycteropus is also diphyodont. Since then other Armadillos have been shown to be diphyodont; and the whole group therefore, so far as concerns those members that have teeth, may in all probability be regarded as typically mammalian in this respect.

These characters are slender enough, but there seem to be no others by means of which the members of this order can be satisfactorily linked together. The fact is, that we have here a polymorphic order which contains in all probability representatives of at least two separate orders. We have at present a very few, and these perhaps highly modified, descendants of a large and diverse group of mammals. For convenience' sake they will be all treated of under the head of Edentata.

Although for the probable reasons already stated it is a hard matter to frame such a definition as will include all existing Edentates, it is easy enough to define two groups in this heterogeneous order; to define one group we should say, rather, and then to regard the leavings as forming another not so easily definable a group.

The perfectly-definable group is that which includes the American Anteaters, the Armadillos, and the Sloths. In all these creatures, which may certainly be regarded as representing on their own account as many family types, there are a number of important and highly-characteristic anatomical features which they share in common. So exceedingly different are these three types in general appearance and (correlated with that) in way of life that these common characters acquire increased importance.

Fig. 89.—Great Anteater (Myrmecophaga jubata). A, Side view of twelfth and thirteenth thoracic vertebrae. B, Posterior surface of second lumbar vertebra. C, Anterior surface of third lumbar vertebra, × ⅔. az, Anterior zygapophysis; az¹, az², az³, additional anterior, articular facets; cc, facet for capitulum of rib; m, metapophysis; pz, posterior zygapophysis; pz¹, pz², pz³, additional posterior articular facets; t, transverse process; tc, facet for articulation of tubercle of rib. (From Flower's Osteology.)

The first of these characters is the series of additional zygapophyses on the posterior dorsal and lumbar vertebrae; these are very clear in the Anteaters and Armadillos; less clear, but still obviously represented, in the Sloths. In the second place, they all possess a clavicle, rudimentary, it is true, in the Great Ant-bear, but still present. Thirdly, the testes are abdominal throughout life, a character which they share with such lowly-organised animals as the Monotremata and the Whales. Finally, and this is by no means a matter to be overlooked, not only are all the existing members of this group American in range, but there is no evidence to prove that they have ever existed elsewhere. No European or Old-World

representatives have as yet been discovered which can be referred to the Anteater, Armadillo, or Sloth type with certainty.[^[94]]

Of these American forms, which will be treated of first, the Armadillos are further apart from either Sloths or Anteaters than the last two are from each other. The name Xenarthra has been suggested for the American Edentates with "abnormal" vertebral articulations; the corresponding Nomarthra includes the Old-World forms.

Fig. 90.—Right scapula and clavicle of Two-toed Sloth (Choloepus hoffmanni). × 1⅔. a, Acromion; af, prescapular fossa; c, coracoid; cl, clavicle; csf, coraco-scapular foramen; gc, glenoid cavity; pf, postscapular fossa. (From Flower's Osteology.)

Between the Sloths and Anteaters the extinct Megatherium and some of its allies are to a certain extent intermediate. But it may be pointed out in the first place that there are certain important resemblances between the living forms. In both, retia mirabilia are developed in the tail (in spite of its reduction in the Sloths) and in the limbs. But, as is well known, retia are also found in other mammals far removed in the series from these under consideration. The reproductive organs generally are very similar, and they have both a dome-shaped and deciduate placenta. The latter character they share with the Armadillos and with the Aard Vark; Manis having a non-deciduate placenta which is, like that of the Carnivora, zonary in form. The Edentates, at any rate the American forms, have a double vena cava posterior and no azygos vein. This condition is also met with among Whales.

Osteologically the Sloths and Anteaters are united by the fact that the coracoid becomes fused with the coracoid border of the scapula, thus forming a foramen; the importance of this character is, however, discounted by its occurrence in three genera of Cebidae.

The above facts embody the views of Sir William Flower.[^[95]]

A subsequent study of the brain and of the muscles of these animals has led to results not entirely in harmony with these views.

Dr. Elliot Smith is of opinion,[^[96]] after an exhaustive study of the Edentate brain, that in this region of the body the present group shows very decided points of likeness to the Carnivora; that is, so far as concerns the Anteaters. On the other hand, Orycteropus is as distinctly comparable with a primitive Ungulate type, such as is exemplified by Moschus. "If the brain of Orycteropus," he remarks, "were given to an anatomist acquainted with all the other variations of the mammalian type of brain, there is probably only one feature which would lead him to hesitate in describing it as an exceedingly simple Ungulate brain. That one feature is the high degree of macrosmatism.[^[97]] Manis, on the other hand, does not come especially near to Orycteropus. The brain of Manis conforms to a simple type of architecture, which agrees in many points with both those of Orycteropus and the American Edentates; there is not sufficient evidence to show which type it really favours." Elliot Smith would, in fact, agree with Max Weber that it is better, if a division is to be made, to divide the group into three orders:—the Xenarthra (Sloths, Anteaters, and Armadillos), Tubulidentata (Orycteropus), and Squamata (Manis), instead of into Xenarthra and Nomarthra.

Messrs. Windle and Parsons[^[98]] are disposed to see in muscular similarities reasons for uniting Manis with the American Edentates, though they confess to being unable to place Orycteropus; in this animal, they say, "we are more struck by the generalised mammalian arrangement of its muscles than by any special Edentate characters. There are, however, two muscles in Orycteropus which show peculiarities not found elsewhere than in the Edentates";—the triceps, which has more than one scapular head, and the tibialis posticus, which is double. They conclude that Orycteropus "presents some feeble claims to be taken into the order."

We shall here adopt the following divisions.

Sub-Order 1. XENARTHRA.

Fam. 1. Myrmecophagidae.—The family Myrmecophagidae contains three genera, all South American in range. These genera, Myrmecophaga, Tamandua, and Cycloturus, agree greatly in their outward form. They are all without teeth, and have long snouts and long protrusible tongues. The fur is thick, and they have powerful claws wherewith to break down the strong ant-hills upon whose inhabitants they feed. Tamandua and Cycloturus are arboreal, Myrmecophaga is terrestrial in habit. The claws of the arboreal forms are useful to destroy the bark, and thus bring to light insects which lurk in such situations.

Fig. 91.—Great Anteater. Myrmecophaga jubata. × 1⁄10.

The genus Myrmecophaga contains but one species, the Great Anteater, Myrmecophaga jubata. It is a large and handsome animal, with long, shaggy, greyish-black hair and a broad white stripe across the shoulder. The coloration is similar in the two sexes. Including the long and bushy tail it reaches a length of over 7 feet. It is on account of its long tongue and greatly developed salivary glands that this and the allied genera were originally placed with Manis. It is the submaxillary glands which are so enormous; they extend back over the chest, and open by three distinct ducts, of which two unite just before the external orifice.

Along their course these ducts are provided with a sphincter muscle, which squeezes the secretion towards the external orifice into the mouth-cavity. The stomach is somewhat gizzard-like. The intestine has no caecum.[^[99]]

The Anteater's great claws are not only serviceable in tearing up the ground to get at its food; armed with them he does not fear, as Mr. Waterton remarked, "the fatal pressure of the serpent's fold or the teeth of the famished jaguar." An Anteater, too, is more than a match for a big dog, and will rip open its belly with the claws while the dog is vainly trying to make an impression with its teeth upon the shaggy hair.

Tamandua is a smaller animal than Myrmecophaga, and, as has been stated, is arboreal; associated with this habit is a prehensile tail. Like the last genus, Tamandua has a rudimentary clavicle, this bone being well developed in the little Cycloturus.

Fig. 92.—Skull of Anteater (Myrmecophaga). Lateral view, al.sph, Alisphenoid; cond, condyle of mandible; cor, coronoid process of mandible; ex.oc, exoccipital; ext.aud, external auditory meatus; fr, frontal; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; occ.cond, occipital condyle; pal, palatine; par, parietal; p.max, premaxilla; s.oc, supraoccipital; sq, squamosal; ty, tympanic. (From Parker and Haswell's Zoology).

The skull of the Anteater[^[100]] is very long and low; the fore-part is tubular, and there appear to be no traces of teeth. The premaxilla is very small; the zygomatic arch is imperfect, and does not reach the squamosal behind. A curious feature of this genus, which it shares with some Dolphins and other Whales, is that the pterygoid bones develop palatine plates which meet each other in the middle line, and thus shift the opening of the

posterior nares backwards. This is also, of course, a character of various lower vertebrates. Another Whale-like character in the skull is the weak character of the mandible, which does not give off a marked coronoid process. But then in neither group is there much mastication. The tympanic, periotic and squamosal are ankylosed together. A peculiarity of the cervical vertebrae is that (as in the Camels) the vertebrarterial canal of several of the vertebrae perforates the pedicle obliquely. There are fifteen or sixteen dorsal and three or two lumbar vertebrae. The additional zygapophyses upon the former have been already referred to. The mode of articulation of the ribs is highly singular.

Each segment of the sternum (of which there are eight) is separated from the next by a synovial membrane: and it has on either side two facets for articulation with the ribs. The way in

which these latter bones are connected with the sternum is curiously like their mode of connexion with the spinal column at their other end. With this may be possibly compared the double articulation of the single rib (which articulates with the sternum) in the Rorquals. In Cycloturus this mode of articulation does not occur.

The manus of Myrmecophaga is five-fingered. Of these the third digit (as in Perissodactyles) is the most prominent; it is at least double the width of the second or third finger; the pollex is very slender. In the little Cycloturus this is carried to a greater extent: the third digit is relatively enormous; the first and the fourth have become quite rudimentary; while the fifth is only just recognisable as a minute ossification.

Fig. 95.—A, Manus of Great Anteater (Myrmecophaga jubata). × ⅓. B, Manus of Little Anteater (Cycloturus didactylus). × 2. c, Cuneiform; l, lunar; m, magnum; p, pisiform; s, scaphoid; td, trapezoid; tm, trapezium; u, unciform; I-V, digits. (From Flower's Osteology.)

The chevron-bones in the tail surround a well-developed rete mirabile, a rete being found in precisely the same position in the Eastern Manis. Tamandua has also retia, which are also found in the Spider-monkeys.

Cycloturus is by far the smallest of the Anteaters. It has

only two toes on the fore-feet. It is to be distinguished, anatomically, from its larger relatives by the complete clavicle, and by the fact that the pterygoids do not meet in the middle line of the skull. The ribs, too, are unusually wide, as in the Whale Neobalaena, and form a bony encasement for the body. It has two small caeca. Of fossil Anteaters but little is known. The most interesting form is Scotaeops, interesting because it has two small back teeth, which are totally lost in its living allies. The huge Patagonian extinct bird Phororhacos, first known by a lower jaw, was at one time regarded as a member of this group on account of the form and edentulous character of the jaw.

Fig. 96.—Unau, or Two-toed Sloth. Choloepus didactylus. × 1⁄5. (After Vogt and Specht.)

Fig. 97.—Skull of Three-toed Sloth. Bradypus tridactylus. Lateral view. fr, Frontal; ju, jugal; lcr, lachrymal; max, maxilla; nas, nasal; par, parietal; s.oc, supra-occipital; ty, tympanic. (From Parker and Haswell's Zoology.)

Fam. 2. Bradypodidae.—The Sloths, genera Bradypus and Choloepus, come, as already stated, very near to the Anteaters, in spite of their striking difference in appearance. The Sloths are purely arboreal creatures, with strong recurved claws, which serve as hooks to keep them suspended from the lower side of a branch. The three-toed sloth, Bradypus (or "Ai"), has the exceptional number of nine cervical vertebrae; the two-toed sloth, Choloepus hoffmanni (or "Unau"), has the equally exceptional number of six. The hair is long and shaggy, and gets an adventitious green colour from the presence of minute algae.[^[101]] This gives to the animal the appearance of a lichen-covered bough, a resemblance which is increased in one species by an oval mark upon the back, which suggests forcibly a broken end of such a branch. The likeness of a Sloth to its surroundings is pointed out by Dr. Siemann,[^[102]] who observed that a species occurring in Nicaragua "has almost exactly the same greyish-green colour as Tillandsia usneoides, the so-called 'Vegetable Horsehair' common in the district.... If it could be shown that it frequented trees covered with that plant ... there would be a curious case of mimicry between the sloth's hair and the Tillandsia, and a good reason why so few of these Sloths are seen." The stomach in the Sloths is complicated in structure, with several chambers; one of these gives off a long crescent-shaped caecum. The skull of the Sloths agrees in a number of particulars with that of the Anteaters.

Fig. 98.—Skeleton of Three-toed Sloth. Bradypus tridactylus. (After de Blainville.)

The zygoma is incomplete, though the part connected with the frontal has a strong downward process like that found in Diprotodon and some other mammals. There is, moreover, a process from the squamosal, though it does not reach the anterior part and thus

complete the arcade. The premaxillaries are very small, and are usually lost in dried skulls. Coupled with these points of likeness are some differences. The lower jaw, for instance, has a well-marked coronoid process. The pterygoids do not meet in the middle line. The teeth are five or four in each half of each jaw. There is no trace of a second set.

A peculiarity of the Sloths is the enormous number of dorsal vertebrae. There are twenty-three of these in Choloepus hoffmanni, but only fifteen to seventeen in the Three-toed Sloth, Bradypus. As in other American Edentates, the acromion joins the coracoid. This connexion occurs in both the Two-toed and the Three-toed species. The limbs of these creatures are very long, a concomitant of an arboreal life. The femur has no third trochanter. The genus Bradypus, which by reason of the fact that it has not lost the third toe on the manus seems to be more primitive than Choloepus, shows another structural feature which does not bear out this conclusion. The trapezoid and the os magnum of the carpus are united, while in Choloepus they are perfectly distinct bones.

The intestine has no caecum.

There are several species of Sloths. Eminently perfect though the organisation of the Sloth in relation to its particular surroundings appears to us, Buffon selected the animal as the very type of imperfection in nature. "One more defect," he wrote, "they could not have existed."

Fam. 3. Dasypodidae.—The family Dasypodidae or Armadillos contains a considerable number of genera. Tatusia, Tolypeutes, Dasypus, Xenurus, Priodon,[^[103]] and Chlamydophorus. They have all a more or less rigid covering of bony plates imbedded in the skin, which are not in the least comparable with the scales of the Manis. Save the Whales, in one or two genera of which traces of a dermal armature exist, the Armadillos are unique among existing mammals in this particular. The term "Edentate" is especially inapplicable to the Armadillos; the genus Priodon may have more than forty teeth in each jaw; a total of ninety was found in one specimen examined by Professor Kükenthal. In the tendency of the teeth to multiply, we have another example of a state of affairs which characterises so many Whales. Generally, however, seven to nine is the number of teeth in each

half jaw, of which one is often implanted in the premaxilla. The Armadillos show their alliance with the other American Edentates in the points enumerated above. Their teeth specially ally them to the Sloths, while the salivary and digestive organs generally are on the Anteater plan, but present a less extreme development. There are, however, caeca, paired as in birds, in the genera Dasypus and Chlamydophorus. The others have none. But there is a dilatation at the commencement of the large intestine, which is not very different from the slightly-developed caeca of Dasypus.

There are certain peculiarities in the skeleton, which distinguish this family.

Fig. 99.—Skull of Armadillo. Dasypus sexcinctus. × ⅔. ex.oc, Exoccipital; fr, frontal; max, maxilla; nas, nasal; par, parietal; peri, periotic; p.max, premaxilla; s.oc, supraoccipital; sq, squamosal; ty, tympanic. (From Parker and Haswell's Zoology.)

The skull in the Armadillos presents a number of likenesses to the other American Edentates.[^[104]] The premaxillaries are small, but are larger in Dasypus than in Tatusia. On the other hand the lachrymals are larger in the latter. The zygomatic arch is complete, but there is no downward process as in the Sloths. In Tatusia (but not in Dasypus) the "short thick pterygoids add somewhat to the hard palate." This is clearly a beginning or a remnant of the quite crocodilian character of the palate of Myrmecophaga. In the cervical vertebrae we see the Whale-like character of fusion between individual vertebrae; and also, as in the Whales, the degree to which this fusion is carried out varies; two to four may be thus united. The additional articular facets upon the dorsal vertebrae have been already commented upon as a point of important likeness to other American Edentates. The dorsal vertebrae are commonly eleven in number, the lumbar being three. But in Priodon the numbers are twelve and two respectively. There are traces to be observed of the double-headed attachment of the ribs to the sternum. The shoulder girdle of the Armadillos is somewhat diverse in form in different genera; the acromion is always large, and is remarkable in Priodon for the fact that the humerus also articulates with it, its extremity being recurved, and forming a socket for this purpose. As in some other Edentates there is a second spine on the scapula behind the first. The clavicle is strong. There is some variation in the form of the manus. It is five-fingered in Dasypus; in Tolypeutes the first digit has vanished; on the other hand, in Priodon, the fifth has become rudimentary

and the third enormously enlarged. This latter fact recalls the arrangement characteristic of Myrmecophaga. The pelvis is greatly attached by the ischium to the vertebral column. The femur has a third trochanter.

The various forms of Armadillos are largely distinguished by the number of movable thin bands of scutes lying between the large anterior and posterior shields. Thus we have Dasypus sexcinctus, Tolypeutes tricinctus, etc.

Fig. 102.—Pelvis and sacrum of Armadillo. Dasypus sexcinctus. ac, Acetabulum; il, ilium; isch, ischium; obt.for, obturator-foramen; pect.tub, pectineal tubercle; pub, pubis. (From Parker and Haswell's Zoology.)

The little Pichi-chago (or, more correctly, Pichy-ciego), Chlamydophorus, which only grows to about 5 inches in length, has no movable bands at all. It is covered with a uniform series of plates, which, moreover, are not discontinuous at the neck. It differs, too, from the prevailing Armadillo-type by the absence of conspicuous external ears. In the anterior part of the body the armature consists of little more than the horny plates, which in other Armadillos overlie the bony dermal plates. In the hinder region the bony plates are strong. In this animal, therefore, we have the dermal armature reduced to a minimum; but it must be noticed that, like the extinct Glyptodons, the armature is continuous and nowhere ringed.

The genus Tolypeutes, of which the best-known species is T. tricinctus, the Apar (there are two other species in the genus), can roll itself up into a ball like the Pill-Millipede (Glomeris), and, protected by its armour, roll away from its enemies like the Arthropod under similar circumstances. This mode of protection, be it observed, is also adopted by the Pangolin and by the

Hedgehog. The genus has only three movable bands. The tail is short, and is covered with large tubercles. This genus is very markedly digitigrade when running.

Fig. 103.—Three-banded Armadillo or Apar. Tolypeutes tricinctus. × ¼.

Fig. 104.—Peludo Armadillo. Dasypus sexcinctus. × ¼. (After Vogt and Specht.)

The Peludo, Dasypus sexcinctus, is, like other Armadillos, an omnivorous creature, and appears to be particularly fond of carrion. It will burrow up to a decaying carcase like the ground-beetles.

Mr. W. H. Hudson has described the way in which this Armadillo will kill a snake by holding it down and literally sawing the reptile in half by help of the sharp and serrated edges of the carapace. Dasypus has a very short tail, which is shielded by distinct rings near the base.

Tatusia novemcincta is a species with nine movable bands. The genus has four teats; the ears are near together. There are no caeca and no azygos lobe to the lung. A species apparently belonging to this genus, but described under the generic names of Cryptophractus and Praopus, is remarkable for the thick covering of hair, not entirely wanting but usually thin in other Armadillos. In this particular species the coat of hair is so thick as to conceal the underlying plates of the carapace. The individual hairs are stiff, and one inch and a half in length.[^[105]]

The genus Xenurus contains several species, the best known of which is inaptly named X. unicinctus. As a matter of fact the characteristic feature of the genus is the existence of twelve or thirteen movable plates between the two ends of the body. X. unicinctus has twelve dorsal and three lumbar vertebrae. This Armadillo, known by the vernacular name of the Cabassou, has one of the most modified hands that are found in the family. The first two digits are slender and elongated; but are quite normal in the number of their phalanges. In the remaining three digits the metacarpal is short and broad, while the proximal phalanx is either suppressed altogether or fused with the metacarpal, the middle phalanx is present but short, while the third phalanx is very large indeed. As in Dasypus, but not as in Tatusia, which is in so many other respects divergent from these genera, the lungs have an azygos lobe. As a small point of difference, tending to show an alliance between the genera Xenurus and Dasypus and their difference from Tatusia, is the deeply-imbedded gall-bladder; this sac is not nearly so deeply plunged into the hepatic tissue in Tatusia. Xenurus has no caecal dilatations. The brain "is intermediate in its form and surface markings between Dasypus and Tolypeutes." The small intestine is nearly eighteen times the length of the large. But these intestinal measurements are not of much avail in this group as marks of affinity, since in three species of Dasypus Garrod gives the following widely-divergent lengths:—D. villosus, 11.5 feet and 1.25; D. minutus

5.1, with a large intestine of no less than 7 feet; D. vellerosus 4.3 and .66.

Priodon is the giant of its race. This Armadillo may reach a length of 3 feet to the base of the tail. The tail is some 20 inches long. The large number of teeth has been already noticed. There are twelve or thirteen bands. Other points in the structure of this genus have already been mentioned, and need not be recapitulated. This Armadillo feeds upon termites and carrion.

Scleropleura is unfortunately but imperfectly known. The single species, named by Milne-Edwards[^[106]] S. bruneti, is apparently a very rare inhabitant of Brazil. It is known by a single skin, which was tanned by the hunter who obtained it. Thus the hair, if any, has dropped out. The plates in the skin are deficient along the back and even upon the top of the head, and are barely represented upon the tail posteriorly. The ears are small and distant from each other. The tail is longish, about one-third of the length of the body. The total length of the creature including the tail is rather more than a foot and a half. The hunter who obtained it regarded it as a hybrid between an Armadillo and an Anteater.

Extinct Xenarthra.—There are a good many extinct forms of Armadillo, apart of course from the Glyptodons. Peltephilus is referred to later (p. [186]). Dasypus was represented by a large form, 6 feet long, with a skull of one foot in length. The genus Eutatus was also large. The carapace was formed of thirty-three distinct bands, of which the last twelve are soldered together, but not fused into a shield as in Dasypus, etc.

An extinct group of American Edentates, termed the Gravigrada,[^[107]] are somewhat intermediate between the Sloths and the Anteaters. A number of the genera are well known from complete skeletons.

One of the typical forms of this group is Mylodon, which, together with its immediate allies, is often placed in a separate family, Mylodontidae.

Mylodon itself was a large creature, as big as a Rhinoceros. It was covered externally by armour in the skin, which did not form a massive armature as in the Glyptodonts, but was in the

form of scattered plates, small and not fused together. The general aspect of the skull is decidedly Sloth-like. As in that animal, the malar bone is bifid posteriorly, and between the bifurcation is embraced the process of the squamosal. This latter is thus more developed than in the Sloth, but there is no actual union between it and the malar. The premaxilla is small. The lower jaw has both coronoid and ascending processes, and is massive. There are five teeth on each side above, and four on each side below, as in the Sloths. There are the normal seven cervical vertebrae and sixteen dorsals. The limbs are not long and slender,but short and strong, the animal having been terrestrial. The fore-feet were five-toed, of which the three inner toes had claws. The hind-feet were only four-toed, and the two inner only were clawed.

Fig. 105.—Mylodon robustus. (Restoration, after Owen.)

Scelidotherium is a genus which is a trifle smaller than the last. It has only four properly-developed toes in the fore-foot, the thumb being rudimentary; of these, the first two bear claws. The hind-feet are also four-toed. Like Mylodon, Scelidotherium is a Pleistocene genus.

Glossotherium has a skull very much like the last two genera; but it is remarkable for the fact that the nostrils instead of being unprotected with bone anteriorly are there closed by a plate of bone formed by the well-developed premaxillae, the nostrils appearing at the sides, and giving the skull a curious likeness to that of a Chelonian. From a series of recent and most important observations it appears to be clear that this genus has survived into quite modern times.[^[108]]

The well-known naturalist of La Plata, Señor Moreno, engaged in studies connected with the political boundary line between Chili and the Argentine, had occasion to visit Consuelo Cove on Last Hope Inlet in Patagonia. Hanging from a tree he noticed a piece of dried skin, which at once struck him as looking more like the remains of a Mylodon than of any living animal. The inhabitants regarded this piece of skin as a great curiosity, but were of opinion that it was the hide of a cow encrusted with pebbles! This fragment from a bygone age was originally described by Professor Ameghino, who had apparently seen some of the bonelets imbedded in it, as Neomylodon listai, "a living representative of the ancient Gravigrade Edentates of Argentina." That this piece of skin is of quite recent date seems to be proved by a number of considerations. In the first place it is covered by long hair of a light yellowish-brown colour; it does not seem likely that hair would preserve its character for geological epochs. The nearest corresponding case is that of the remains of Moas in New Zealand, whose feathers, dried skin, and tendons are known. Now the Moa was unquestionably contemporaneous with man, as abundant surviving legends prove, and indeed it cannot have been long extinct. Still, hair is a resisting structure, and in a dry cave, with no possibility of irruptions of floods, might retain its characters for long periods. The evidence, however, of more recent date is stronger than this. The skin shows patches of reddish colour, suggestive of course of blood-stains. A small piece of the outside of the skin at the cut edge, which presented the appearance of freshly or comparatively freshly dried fluid, was submitted to a chemical examination and shown to be serum! Dr. Lönnberg examined chemically a bit of the skin itself and found in it, after boiling, glue, "which proves that the collagen and gelatinous substances are perfectly preserved." After this it seems impossible to suppose that the skin can be of any very great age; for bacteria would have finished their work upon the serum and gelatine long ago. Combined with the fresh appearance of the skin is the very fresh appearance of the skull. In fact it is impossible to believe that the animal was not alive quite a few years since, relatively speaking. It is admitted that this animal was contemporaneous with man. There are actually legends of a creature which may have been this Glossotherium. "Ancient chroniclers inform us that the indigenous inhabitants recorded the existence of a

strange, huge, ugly monster, which had its abode in the Cordillera to the south of latitude 37. The Tehuelches and the Gennakens have mentioned similar animals to me, of whose existence their ancestors had transmitted the remembrance; and in the neighbourhood of Rio Negro, the aged Cacique Sinchel, in 1875, pointed out to me a cave, the supposed lair of one of these monsters, called 'Ellengassen'; but I must add that none of the many Indians with whom I have conversed in Patagonia have ever referred to the actual existence of animals to which we can attribute the skin in question."

A rude painting in a cavern, in red ochre, seems to Dr. Moreno (whose words we have just quoted) to be somewhat suggestive of a Glyptodon. There are some reasons for believing that this quadruped was kept by man as a domestic creature. In the cave are two walls of rough pieces of stone which seem to have dropped down owing to the wearing away of the roof; they also seem to have been loosely piled together to form two walls, within which enclosure an imperfect skull of the animal was found. This skull shows clearly that the so-called "Neomylodon" must be referred to Glossotherium or Grypotherium, as it is sometimes termed. This skull is perforated on the roof in such a way as could only have been effected (in the opinion of experts) by a weapon in the hand of a man. A hole in the skin has been even compared to a bullet-wound. But this it is perhaps unnecessary to discuss. The skin of Glossotherium is, like that of other extinct "Ground-sloths" (e.g. Mylodon), filled with small and irregular ossicles. But in Mylodon, the sculptured appearance of the dermal ossicles appears to indicate that they reached the surface of the body and were covered by epidermis alone, which is not the case with the animal now under consideration. The microscopic characters of the ossicles, too, show differences in the two. Glossotherium being "precisely intermediate between Mylodon and the existing Armadillo (Dasypus)." Now Glossotherium and Mylodon are regarded as forms which lie between the existing Anteaters and the Sloths of the same part of the world. We have already pointed out the facts of structure which lead to this conclusion. It might therefore be reasonably surmised that the hair of Glossotherium would be also intermediate, or at least like that of one of the two genera Myrmecophaga and Bradypus. But microscopical investigation has

negatived this supposition. It has shown that the Armadillos are in this matter the nearest relatives of Glossotherium. This result is important as tending further to confirm the close interrelationship of all the American Edentates as contrasted with the Old-World forms—a matter which has already been emphasised. It is suggested, however, that the absence of under fur, which is so well developed in the Sloth, and the difference shown in transverse sections from the hair of Myrmecophaga, may be explained by difference in habitat. Glossotherium lived under conditions similar to those under which the Armadillos live to-day. Thus the outer covering of the body became alike in the two cases, the same needs supervening in both genera.

Lestodon is another allied genus, which seems to possess canines. At any rate, in front of the four molars, and separated from them by a diastema, is a smallish, somewhat canine-like tooth, in both jaws.

Megalonyx and its allies are sometimes placed in a distinct family, Megalonychidae. Megalonyx itself had a skull very like that of Bradypus, being shorter and not so elongated as in the Mylodontidae. There is a strong tusk anteriorly, which is separated by a considerable space from the three molars lying behind it. Both pairs of limbs seem to have possessed five toes. This is a North American genus. It differs from the bulk of the American Edentates in having a complete jugal arch.

Megatherium is the type of yet a third family, Megatheriidae, of the Gravigrade Edentates. This creature is familiar from the many restorations which have been built up, and from its huge bulk, little short of that of an elephant. The skull, which is small for the size of the creature, has a complete jugal arch, from the middle of which depends a downward process as in other allied forms. The teeth grow to an extraordinary depth, and there are five of them in the upper and four in the lower jaw—on each side of course. The fore-limbs of the Megatherium are very much more slender than the enormously bulky hind-limbs, upon which and the equally massive tail the animal seems to have supported itself while tearing down branches of trees, upon whose leaves it fed. In the scapula the acromion joins the coracoid as in Bradypus; the clavicle is large. The fore-limb is four-toed, and the hind-limb three-toed. The latter has but one clawed digit (the third, i.e. the inner).

On the manus, the three inner digits have powerful claws. This animal, too, was Pleistocene in time. The Megatheriidae had, however, small as well as gigantic forms.

The genus Zamicrus had a skull no bigger than that of a Sloth, while Nothrotherium was also a comparatively small creature; the teeth of the latter genus are reduced to 4/3.

The extinct group of the Glyptodontidae comprises large creatures with a dense covering of bony scutes which are arranged in a tesselated fashion, and thus form an immobile armature of immense strength. In correspondence with this massive carapace the dorsal vertebrae have fused together, and the lumbar vertebrae form a series ankylosed to each other and to the following sacrals. These creatures are all South American.

Fig. 106.—Glyptodon clavipes. × 1⁄12. (After Owen.)

Glyptodon, the genus which gives its name to the family, is known from numerous remains in South America, and also from so far north as Texas and Mexico. It grew to be as long as 16 or 17 feet. In the skull there is an exceedingly long downward process of the zygomatic arch, as in Sloths, the arch itself being complete. The process extends so far down as to reach a point about on a level with the middle of the lower jaw. The nasals are short or rudimentary. As in Myrmecophaga, the pterygoids enter into the formation of the bony palate. The lower jaw has a spout-shaped extremity, and, behind, it rises into an enormous vertical branch as high as the front part of the jaw is long. There are eight teeth in each half of each jaw. As in

some Armadillos, the cervical vertebrae are at least partly fused. The atlas is free, but the rest, or at any rate five of them, are united. The last cervical is sometimes fused with the succeeding dorsals; the latter are twelve in number, and are fused together so far as concerns their centra and neural processes. The succeeding region of the vertebral column includes seven to nine lumbars, which are fused with the eight sacrals; in this region the neural processes are high, and there is thus produced a strong and lofty ridge along the back, which forms a powerful support for the carapace. The fore-limbs are shorter than the hind-limbs, which latter are attached to an unusually massive pelvis. The claws of the limbs are blunt and almost hoof-like.

The heavy carapace consists of sculptured, five or six-sided plates, which have no particular arrangement in the middle, but towards the margins show indications of an arrangement in transverse rows. The moderately long tail is also encircled by bony skin-plates which are thorny above, or at least provided each with a blunt upstanding process. It appears that outside this bony system of scutes were horny epidermic scales, corresponding exactly with the tesserae which they cover. There are apparently a good many species of Glyptodon.

In the allied genus Panochthus the tail is rather longer, and the bony rings which surround it, instead of being all movable as in Glyptodon, are at first so, but later, i.e. towards the end of the tail, become welded into a single and massive piece. Both feet are here four-toed, while in Glyptodon the hind-feet are five-toed and the fore-feet four-toed.

Daedicurus shows a further specialisation, in that the feet have three and four digits respectively. The orbit too shows a specialisation in being separated from the temporal fossa. The descending process of the zygomatic arch is not so extraordinarily exaggerated as it is in Glyptodon. It has the same terminal tube of osseous scutes upon the tail. This creature seems to have reached a length of about twelve feet.

Propalaeohoplophorus is, unlike the great Armadillos that we have hitherto dealt with, a small animal, not exceeding 2 feet or so in length of carapace. A small alveolus on each side of the premaxillae seems to suggest the former presence of an incisor tooth; and it seems that the animal possesses both true molars and premolars; for the first four of the eight teeth are much

simpler in structure than those which follow. The dorsal vertebrae again are not fused together; the hind-limbs are five-toed. All the plates of the carapace are arranged in definite transverse rows; it has been observed, too, that some of the anterior scutes overlap like those of the Armadillos, to which this animal possesses further likenesses in the exclusion of the maxillae from the border of the nostril (a Glyptodont character), and the comparative feebleness of the scutes.

A primitive genus also appears to be Peltephilus, which is perhaps rather an Armadillo than a Glyptodon. However, it comes somewhat between the two, like Propalaeohoplophorus, with which it may therefore be treated. A most singular feature of this genus has been mentioned on p. [27] in connexion with the skull in the Mammalia generally. That is the fact that a portion of the squamosal surrounding the articular facet for the lower jaw is separated by a suture from the rest of that bone, and is therefore obviously suggestive of the quadrate in the lower Vertebrates. As in certain Armadillos and Glyptodons, etc., the pterygoids appear in this genus to have taken a share in the formation of the hard palate. The plates of the carapace were movable, as is shown by the fact that they sometimes slightly overlap. In view of the possible origin of the Edentates from lowly-organised Mammalia, it is noteworthy that the humerus has been especially compared to that of the Monotreme. Peltephilus differs from other Armadillos in having teeth in the front of the jaws. The total number of teeth is twenty-eight, i.e. seven in each half of each jaw.

Sub-Order 2. NOMARTHRA.

As already explained, the Old-World Edentates differ from the New-World forms in having normal dorsal vertebrae, that is to say, without additional zygapophyses. That negative feature, however, though combined with the positive fact that both the Old-World forms feed upon ants, is hardly sufficient to outweigh the many structural differences which distinguish the Orycteropodidae from the Manidae; which will be placed therefore in different groups. To that containing the Aard Vark, the name Tubulidentata may be applied.

This group contains but one family, the Orycteropodidae, of which there is but a single genus.

The Aard Vark (earth-pig), genus Orycteropus, is characterised by its heavy build, the body being covered by rather coarse and not very abundant hair; the snout is long and pig-like, with round nostrils at its end; the ears are long, erect, and pointed; the tail is very thick at first, so that it has been aptly described as "a tapering of the body to a point." The fore-limbs are four-toed, the hind five-toed.

Fig. 107.—Aard Vark, or Cape Anteater. Orycteropus capensis. × 1⁄16.

In the skull there is a complete though slender zygoma; the premaxillaries, though small, are not so rudimentary as in the American Edentates. The annular tympanic is not ankylosed to the surrounding bones, a character found in other low mammals. Contrary to what is found in Manis, Orycteropus has a huge lachrymal. There are thirteen dorsal and seven lumbar vertebrae. The clavicle is well developed. Orycteropus is peculiar among Edentates in that the ischia do not unite with the vertebral column. The femur has a third trochanter.

As mentioned on p. [162], the Aard Vark is diphyodont like normal mammals. The permanent teeth consist of five molars and premolars on each side of each jaw; the first two of these are premolars, and are simpler in their form than the succeeding two teeth, which are partly divided by a median furrow into two halves. These teeth are also peculiar in that they consist entirely of vaso-dentine. They have been compared in minute structure to those of the Ray Myliobates. According to Mr. Oldfield

Thomas[^[109]] there are seven milk teeth on each side of the upper jaw (limited to the maxillae, and thus not incisors). An eighth tooth was discovered on one side of one of the specimens examined by Thomas. In the lower jaw there are only four milk teeth on each side. It is interesting to note that the histological structure of these milk teeth agrees with that of the permanent teeth. There are two species of this genus found in Africa: the southern, O. capensis, is more hairy than the northern, O. aethiopicus. O. gaudryi is a Pliocene species from the Island of Samos and from Persia, described by Dr. Forsyth Major and Dr. Andrews.[^[110]] It closely resembles the existing O. aethiopicus.

Fig. 108.—Section of lower jaw with the teeth of Orycteropus. × 2. (After Owen.)

Of the Scaly Anteaters, Group Squamata or Manidae, there is really but one genus, though Phatagin, Pholidotus, Smutsia, and Pangolin have been used to distinguish various forms. The genus Manis is African and Oriental in range. Dr. Jentink, who has lately revised the species, allows seven.[^[111]] The external form of these animals is fairly well known, the remarkable scales distinguishing the Pangolins from other animals. Between the scales lie hairs, which seem to be absent in the adults of the African species, though present in the young, thus affording a convenient method of distinguishing the Ethiopian from the Oriental forms. The scales have been compared to agglutinated hairs. That they are not "merely mimetic of the Lizards' scales" is held by Weber,[^[112]] who compares them directly with those

structures, as he does the scales of other mammals, such as those upon the tail of Anomalurus, etc. This, however, is not a universal opinion. It is true that these scales occur chiefly in the lower forms of mammals such as those under consideration, Marsupials, Rodents, and Insectivores; but the fact that the hairs are developed before the scales shows, or seems to show, that the former are the older structures, and to lead to the inference that the scales of mammals are new structures. The scattered hairs of the Pangolin have no sebaceous glands excepting on the snout. This, again, looks as if they were degenerate structures, and emphasises the non-archaic character of the scales. These animals have no trace of teeth except possibly some slight epithelial thickenings which have been interpreted as a last remnant; the tongue is suited for the capture of ants, and is therefore much like that of the not nearly-related American Anteaters. The stomach is of simple form; it is characterised by a large gland, which suggests that of the Koala (see p. [144]); the intestine has no caecum. Retia mirabilia occur on the limb arteries. The placenta is non-deciduate and diffuse; it is specially compared by Weber with that of the Horse. Considering the many adaptive resemblances between this genus and the American Anteaters, especially in the mouth cavity, it is remarkable that in Manis the pterygoids are not joined as they are in Myrmecophaga. In spite of statements to the contrary, it appears that there is sometimes a distinct lachrymal.

A remarkable feature in the skeleton of Manis is the singular sternum. The xiphoid cartilage is extraordinarily elongated into thin strips, which reach the pelvis and return. This state of affairs is to be found in the African species only. This structure is not comparable, as it has been said to be, with abdominal ribs such as those of the reptile Hatteria.

These animals are mainly anteaters. The Japanese have a curious legend as to the method adopted for the capture of ants, which is related by Dr. Jentink in his monograph of the genus. The Manis "erects his scales and feigns to be dead; the ants creep between the erected scales, after which the anteater again closes its scales and enters the water; he now again erects the scales, the ants are set floating, and are then swallowed by the anteaters"! The same story is related by Mr. Stanley Flower on the authority of the Malays.

Though it seems clear that the likenesses which Manis shows

to the Anteaters of the New World are chiefly adaptive and have nothing to do with real affinity, being merely an expression of a similar mode of life, it is curious to note that here and there we do find certain resemblances which do not seem to be susceptible of the latter explanation. The jugal bone, absent in Manis, is small in Myrmecophaga; the clavicle is absent and again small or rudimentary in the Anteaters; it is large in other Edentates. The third trochanter is absent, as in Myrmecophaga (and the Sloths). There are many scales on the body; in Myrmecophaga there are traces of these structures on the tail, as also in Tamandua. In the features mentioned, the Myrmecophagidae differ from either or from both of the two other American families (i.e. Dasypodidae, Bradypodidae) and agree with Manis. The facts are not a little remarkable.

Fig. 109.—Manis. Manis gigantea. × 1⁄12.

Order III. GANODONTA.[^[113]]

Allied to the Edentata, and apparently representing the ancestral forms from which they, at any rate the Xenarthra were derived, is the order of the Ganodonta. Of this order a number of genera are now known, which can be ranged in a series which more and more approaches the Edentata as we pass from the older to the newer forms. This interesting and transitional series will be made manifest by a description of the characters of the various genera taken in their proper

chronological order. The following genera are included by Wortman in his family Stylinodontidae.

The earliest type of the Ganodonta is the genus Hemiganus, with but one species, H. otariidens. This animal lived during the deposition of the lowest Eocene strata, the Puerco beds of North America. It was about as big as a fair-sized Dog, and had powerful jaws. There were at least two pairs of incisors in the upper jaw, together with powerful canines and the full premolar and molar formula. In the lower jaw the canines were also strong, but the incisors are not certainly known to be more than two pairs. The enamel upon the posterior surface of the canine is thin, and in the case of the incisors the enamel seems to be limited to the anterior face. The lower molars are quadritubercular. It is believed from the presence of a suture on the upper surface of the premaxillary that the snout of the creature was tubular. The cervical vertebrae, only known by their centra, are like those of the Armadillos (and for the matter of that of the Whales) in the great transverse as opposed to the antero-posterior diameter. The feet are especially compared with those of the Ground Sloths. The single ungual phalanx is marked by a large subungual process, which is pierced by a considerable foramen. The tibia again is to be compared with that of the Armadillos.

In the Upper Puerco (Torrejon) beds the remains of Psittacotherium are found. This genus, when first discovered, was referred to the Tillodontia by some and to the Ungulates, the latter being a refuge for indeterminate Eocene mammals, just as the "Multituberculata" is for similarly-placed Secondary mammals. It is now known to be clearly a member of the order Ganodonta. Wortman thinks that there is but one species, P. multifragum. It seems to have had a general aspect much like that of Hemiganus—that is judging from the skull—and was not very greatly different in size. The facial portion of the skull is short, and the zygoma is deep. The infra-orbital canal is double, a feature which crops up in the Sloth, and has been mentioned in the later form of Ground Sloth, Megalonyx (but it must be remembered that the same characteristic is not unknown in Rodents). The dentition is reduced as compared with that of Hemiganus, that is to say, as far as concerns the molars and the incisors. There is but a single pair of incisors in each jaw; the canines are strong; the premolar and molar series seem to have been complete in the lower jaw,

but reduced by one premolar at least in the upper jaw. It is very important to notice that the incisors have enamel only on their anterior faces, and that the same is the case with the canines, the slender layer present behind the tooth in Hemiganus having vanished in this later form. The tooth pattern of the molars is like that of Hemiganus. The fore-limb is decidedly Edentate-like; but it is the foot which presents the strongest likenesses to that order. "If an anatomist," remarks Dr. Wortman, "had no other part of the skeleton than that of the foot to guide his judgment, and he should fail to detect a most striking similarity between it and that of the Edentata, especially the Ground Sloths, he would not only lay himself open to the criticism of being lacking in the ordinary powers of observation and comparison, but would be suspected of placing the matter upon a basis other than that established by such a method." It is not certain how many toes upon the fore-limbs were possessed by Psittacotherium, but the close resemblance to Mylodon is indeed striking, the third digit being in both forms the most pronounced. Some vertebrae of this Ganodont have been discovered which do not show the complex articular arrangements of later American Edentates. The sacrum, on the other hand, is very like that of the Sloth, and there is a foreshadowing of the attachment of the ilia to the sacrum by co-ossification which is met with in later Edentates. A still later type is the genus Calamodon, which has been shown to occur in Europe as well as in America. C. simplex was a larger beast than either of the genera that have already been treated of, thus affording another example of the increase in size of later as compared with earlier members of the same group, so pronounced among the Ungulata. The lower jaw has the same massive structure that characterises that bone in Hemiganus and Psittacotherium. There is but one incisor, but the premolar and molar series are complete. The canine is Rodent-like in appearance, being imbedded throughout the greater part of the lower jaw; it evidently grew from a persistent pulp. It is enamelled upon the anterior face only. The premolar and molar teeth are in this genus commencing to lose their enamel, which is distributed in the form of vertical bands, leaving interspaces which are not covered by enamel. These teeth, moreover, are decidedly hypselodont, more decidedly so than in Psittacotherium; they are, when unworn, quadricuspidate, with accessory cusps; when more worn, the teeth

are double-ridged, and that transversely to the long axis of the jaw; finally, the much-worn teeth have flattish crowns more or less surrounded by a ring of enamel.

A still later form, coming from the Lower and Middle Eocene strata, is the genus Stylinodon. S. cylindrifer, which is the more archaic of the two described species, is only known from a single molar, fragments of a canine, and "some inconsiderable pieces of the skull." The molar is interesting on account of the fact that the enamel is still further reduced; it is represented only by narrow vertical strips, which are much narrower than those of older forms of Ganodonts. It is also hypselodont, and has a persistent pulp. So, too, the canine which had a thick anterior facing of enamel. The later species, S. mirus, is more fully known. The teeth seem to have been much the same as in the last-described species; the premolars and molars were seven in all in the lower jaw, and the canine was imbedded in the bone for a long distance, as in Calamodon. The cervical vertebrae have short centra as in Hemiganus. The clavicles were well developed. The humerus possessed an entepicondylar foramen, and its head displays the pyriform pattern so characteristic of later Edentates. The foot is clearly like that of Psittacotherium.

In reviewing the series, therefore, we see a gradual diminution of the incisors, a gradual loss of enamel on the teeth generally, and the production of hypselodont teeth growing from persistent pulps; all of which are features of the later Edentates. The progression is so gradual that the forms enumerated and described seem to have been part of a continuous series culminating in the Ground Sloths of later times. The other points of similarity will be gathered from the facts given in the foregoing pages.

There is another family belonging to the Ganodonta whose position with regard to the Edentata is not so clear. This is the family Conoryctidae, of which two genera are known. The earliest of these, from the Lower Puerco, is Onychodectes. In O. tissonensis the skull is long and narrow, thus contrasting with that of the last family. The facial part is also long. The lower jaw is much more slender. The molar formula was complete, but there is some doubt as to the incisors. The molars are tritubercular.

The other known genus is Conoryctes. Its skull has a shorter

facial portion, and is thus more like that of Stylinodontidae than that of Onychodectes. The dental formula is known, and is complete save for the loss of one incisor above and below, and one premolar above. The relationship of these Ganodonts to any later forms is uncertain; but their skeletal structure is as yet by no means fully known.

CHAPTER IX

UNGULATA—CONDYLARTHRA—AMBLYPODA—ANCYLOPODA—TYPOTHERIA—TOXODONTIA—PROBOSCIDEA—HYRACOIDEA

Order IV. UNGULATA

The existing members of this order can be readily grouped into the Hyracoidea, Proboscidea, Perissodactyla, and Artiodactyla, each of which divisions has quite the value of an order, and all of which are sharply marked off from each other. But as the discovery of so many fossil forms has to a great extent rendered these demarcations less sharp, it is better to regard all these groups as not more than sub-orders of a larger "Order" Ungulata. Even when this conclusion has been necessarily arrived at from a consideration of the more ancient groups of Ungulate animals, the definition of such an order remains a difficult matter for the systematist. For the earliest of these forms, more particularly the Ancylopoda, the Amblypoda, and the Condylarthra, whose peculiarities will be dealt with at length subsequently, are not by any means easily differentiated from the primitive Carnivorous mammals of that date, the Creodonta; these latter, moreover, fade into the Marsupials through the so-called Sparassodonta of Professor Ameghino. To confine ourselves to the Ungulates, we may perhaps define them as terrestrial animals with hoofs rather than claws or nails, and chiefly, if not entirely, vegetarian in habit. The teeth are bunodont or lophodont, the tendency to the production of the latter type being always marked. The walk, although plantigrade in the older types, becomes more and more digitigrade, except in such survivals from antiquity as Hyrax. There is, too, as we pass from the ancient types to the modern, a gradual perfection of the limbs as running

and not climbing or grasping organs; the number of toes becomes reduced, and culminates twice (in the horse and in the Litopterna) in one toe on each foot; at the same time the ulna becomes rudimentary and fuses with the radius, and the fibula in the hind-limb undergoes a like reduction. The clavicle is absent even in some of the oldest types; its presence in Typotherium[^[114]] is highly remarkable. The tail too, an organ which is long in some of the early forms, gets short in their modern derivatives.

Fig. 110.—An early Ungulate. Phenacodus primaevus. × 1⁄12. (After Osborn.)

Fig. 111.—Series of metacarpals and metatarsals of Camelidae, to show secular and progressive increase in size. From left to right the species are Protylopus petersoni, Poebrotherium labiatum, Gomphotherium sternbergi, Procamelus occidentalis. F, Fore-foot; H, hind-foot; III, IV, third and fourth metapodials. (After Wortman.)

Coupled with the increasing perfection of the foot as an organ used merely for the support of the body, certain interesting changes have taken place in the arrangement with regard to each other of the several bonelets of the wrist and ankle. It has been held by Cope and others that the truly primitive disposition of these bones was that presented to us by certain early types, such as Meniscotherium or the existing elephant or Hyrax. In these animals there is (see Fig. 112) a serial

arrangement of these bones, the distal bones only, or very nearly only, articulating with the corresponding bones in the upper series. In the modern types (cf. Fig. 113) there is, on the other hand, an interlocking, so that the bones of the distal series articulate with two of those of the proximal series. By this is produced, as it would appear, a much firmer foot, less liable to "give" under pressure, and thus more fitted for an animal that runs. It is the same principle as that adopted in the laying of bricks. The actual stress and strain of impact has been held responsible for those changes. An equally ingenious and possibly truer explanation of the undoubted facts has lately been advanced by Mr. W. D.

Matthew.[^[115]] He has pointed out that in some ancient Ungulates the carpus is not serial but interlocking, even in forms which belong to the earliest Eocene groups, such as the genus Protolambda among the Amblypoda. Now in the fore-foot of Meniscotherium and the living Hyrax there is a separate centrale which is wanting in the greater number of Ungulates. The absorption, that is the practical dropping out of this bone, would restore to an interlocking carpus the serial arrangement; while on the other hand, by the fusion of this bone with the scaphoid, the interlocking disposition would be maintained.

Fig. 112.—Bones of the manus A, of the Indian Elephant, Elephas indicus. × ⅛. B, of the Cape Hyrax, Hyrax capensis. × 1. c, Cuneiform; cc, centrale; l, lunar; m, magnum; p, pisiform; R, radius; td, trapezoid; tm, trapezium; s, scaphoid; u, unciform; U, ulna. (From Flower's Osteology.)

Fig. 113.—Bones of the manus A, of Rhinoceros, Rhinoceros sumatrensis. × 1⁄5. B, of Pig, Sus scrofa. × ⅓. Letters as in Fig. 112. (From Flower's Osteology.)

The gradual perfecting of the fore- and hind-limbs as running organs has been put down to the advent of the grasses, and the formation of large plains covered with this herbage. The same reason would also be in harmony with the equally gradual change in the shape of the molar teeth, from a tubercular form calculated for a mixed or even a carnivorous diet, to the flatter crushing surfaces exhibited by the lophodont teeth of later Ungulates. Strong

canines would in the same way cease to be useful, and even become encumbrances to such grazing creatures; and their disappearance is one of the salient features in the history of the Ungulata, that is of the modern representatives of the order. The extraordinary hypertrophy of these teeth in such a line as that of the Amblypoda, which has left no descendants, was one of the reasons perhaps for the decay of those great pachyderms of mid-Tertiary times; their excessive armature became an encumbrance, since it was not accompanied by improvements in other necessary directions. Some of the features of the Tertiary Ungulates have, however, been dealt with in our general sketch of the mammalian life during that epoch, and need not be again referred to here. Of existing Ungulates there are no clear indications of the descent of the Elephants or of the Hyracoidea. Their structure proclaims these two divisions to be of ancient descent, and not to be modern twigs of the Ungulate stem. As to the Perissodactyla and the Artiodactyla we cannot bring them together nearer than in quite early Tertiary times. The order Condylarthra seems to be the starting-point of both these sub-divisions. Euprotogonia has been considered to be an ancestor of the Perissodactyle branch, and Protogonodon or Protoselene of the Artiodactyla. If this be true,

the likenesses which Titanotherium shows to the Artiodactyla must be either purely superficial and secondary, or a cropping out of ancient characters which had been dormant for many generations.

Horns.—The Ungulata are the only order of mammals which possess horns; as they are on the whole a more defenceless group than the Carnivora, it may be that the horns are a counterpoise to the teeth and claws of the latter; need for defence and for armature in the combats with their own kind for the favours of the does has led to a different kind of protective and aggressive mechanism. Horns as weapons are, however, particularly effective in this group wherever they exist. A Ruminant is most frequently a large and heavy animal without the agility and litheness of the Carnivore. It is precisely to this sort of animal, where weight is an important consideration, that horns are the most suitable weapons. This is further shown by the fact that although the general term horn is used to describe the weapons of the Ungulate mammals, there is more than one kind of structure included under this general term; it is indeed probable that the extreme terms in the series of horns have been independently acquired by their possessors. There is but little in common between the horns of a Giraffe and of a Rhinoceros. In the Rhinoceros we have one or two horns, in the latter case one placed behind the other, which are purely epidermic growths; they may indeed be regarded as matted masses of hair, borne, it is true, upon a boss of bone, which however is not a separate structure. The Giraffe supplies us with the simplest term in that series of horns which are partly epidermal and partly bony. The paired horns of this animal have often been contrasted with those of the Deer, for example; but there is no fundamental difference between them. In the Giraffe a pair of bony outgrowths, originally separate from the skull which bears them, but ultimately ankylosed to it, are covered by a layer of entirely unmodified skin. A distinction of undoubtedly practical importance is usually drawn between the Hollow-horned Ruminants, i.e. Oxen, Goats and Antelopes, and the Deer tribe. There is nevertheless no fundamental distinction. In the Antelopes there is a core of bone, the "os cornu" as it has been termed, which is covered by a horny layer, the horn proper, variously modified in shape and size according to the genus or species. In the Deer there is the

same os cornu, which may however be branched, but which is in the same way covered by a layer of modified integument; this is known as the "velvet"; it only lasts for a certain period, and is then torn off by the exertions of the animal itself, leaving behind the bony core, which is popularly termed the horn. It will be clear that here is only a difference of comparative unimportance; the same essential features are present in both groups of animals, but the modification of the epidermis has progressed along different lines. Both can be referred back to the primitive conditions seen in the paired horns of the Giraffe. Even the difference, such as it is, is bridged over by the Antelope Antilocapra, where the os cornu is bifid and the horn is periodically shed, as is the velvet of the stag; but in the stag the bony part of the horn is also shed, a state of affairs which has no parallel in the Hollow-horned Ruminants. The great Sivatherium may conceivably be an annectant form between the two types of compound horns, i.e. those of the Antelope and those of the Deer. This creature had two pairs of horns, of which, naturally, only the bony cores remain; the hinder pair of these were branched. But although so far they resemble the Deer's horns rather than the Antelope's, Dr. Murie has thought that they were covered by a horny sheath and not by soft skin as in the Deer. In any case these horns were apparently never shed, which is a point of likeness with the Antelope and of difference from the Deer. Apart therefore from the nature of the covering of the bony cores, there are good grounds for looking upon them as intermediate between those of the Deer and those of the Antelopes.

The horns of the Ruminants are frequently a secondary sexual character; this is especially the case with the Deer. The Reindeer is, however, an exception, both the stags and the does having horns. That they are associated with the reproductive function is shown by their being shed after the period of rut, the destruction of the velvet at that period, and also by the effect upon the horns which any injury to the reproductive glands produces. Some useful facts upon this latter head have been amassed by Dr. G. H. Fowler,[^[116]] who noticed in a series of stags, horns showing various degrees of degeneration in the antlers produced by varying degrees and periods of gelding. From the facts

here collected it is clear that a direct effect is produced. If we are to regard horns as secondary sexual appendages which have been subsequently handed on to the female by heredity, we should expect to meet with examples of animals now horned in both sexes, of which the earlier representatives had the horns confined to one sex. This is most interestingly shown by the extinct and Miocene Giraffe, Samotherium, of which the male alone had a pair of short horns, while the skull of the female was entirely hornless; the modern Giraffa, as is well known, has horns in both sexes.

It is interesting to note that the existing Perissodactyles and Artiodactyles are to be distinguished by their unpaired or paired horns. But while there are no Artiodactyles with unpaired horns (save occasional sports) the Perissodactyles have more than once tried, so to speak, paired horns, which ultimately proved fatal to them. The Rhinoceros Diceratherium apparently inherited and improved upon the small paired horns of Aceratherium, but it has left no descendant. The paired horned Titanotheria offer another instance of the same apparent incompatibility between the Perissodactyle structure and the persistence of paired horns.

Sub-Order 1. CONDYLARTHRA.

This group is characterised by the following assemblage of characters. Extinct, often plantigrade Ungulates, with five-toed limbs. Bones of carpus and tarsus not always interlocking, but sometimes lying above each other in corresponding positions. The humerus has an entepicondylar foramen. Dental formula quite complete; the molars brachyodont and bunodont. The premolars are simpler than the molars. The canines are small. As with other early types, the zygapophyses are flat and do not interlock. The astragalus is like that of the Creodonta. This group was American and European in range, the remains of its rather numerous genera being of Eocene time. The best-known genus is Phenacodus, of which some account will be given before discussing the, in many cases, more fragmentary remains of other allied forms.

The genus Phenacodus was first described so long ago as 1872, from a few scattered teeth. Since then several nearly complete skeletons have been obtained, and we are in full possession of

the details of its osteology. It was not a large creature (see Fig. 110, p. [196]), about 6 feet in length, with a small head. The feet were more or less plantigrade, and five-toed. The last phalanges of the toes show that they carried hoofs and not claws; yet the fore-feet look a little as if they could be used as grasping organs. The third digit of both hind- and fore-feet exceeds the others, and thus a Perissodactyle-like foot characterised this Eocene creature. The tail is exceedingly long, and must have reached the ground as the animal walked. This is of course by no means an Ungulate character. Still, in the totality of its organisation the animal was decidedly Ungulate, though Professor Cope spoke of Phenacodus as not merely an ancestral Ungulate but as the parent form of Insectivores, Carnivores, Lemurs, Monkeys, and Man himself! The scapula indeed is from its breadth and oval contour rather like that of a Carnivore. The clavicles as in other Ungulates are absent. The femur is Perissodactyle rather than Artiodactyle in the presence of a third trochanter. The creature had fifteen pairs of ribs and five or six lumbar vertebrae. The two bones of the leg which lie below the femur are perfectly distinct and separate. A cast of the brain-case shows that the cerebral hemispheres were smooth and small, the cerebellum of course completely uncovered and nearly as large as the cerebrum. The olfactory lobes were also large. The complete skeleton of Phenacodus has lately been excavated more fully from the enveloping matrix by Professor Osborn,[^[117]] and mounted in what is regarded as the natural position of the beast. It appears that though five-toed it went upon the three middle toes only, and furthermore that of these the middle one was the more prevailing, so that Phenacodus was distinctly "Perissodactyle," at least in habit. Moreover its "long hind-quarters, the long powerful tail ... are reminiscent of Creodont ancestry." The genus was European and American in range.

Meniscotherium ( = Hyracops[^[118]]) comprises several forms of about the size of a fox; they are both European and American in range. The teeth are more distinctly Ungulate in form than those of Phenacodus, with a

-shaped outer wall. The skull is described as possessing "indifferent, primitive characters," permitting a comparison with those of Opossums, Insectivores, and

Creodonta. It has, as in Phenacodus, no orbital ring. The humerus resembles that of a Carnivore rather than that of an Ungulate. The carpus and tarsus are serial. The fibula articulates with both the calcaneum and the astragalus, which is not the case with Phenacodus. It is suggested that these animals are ancestral forms of the Chalicotheres. In the brain the hemispheres do not cover the cerebellum.

More primitive apparently than Phenacodus was the less-known genus Euprotogonia, or Protogonia[^[119]] as it has been called. The best-known species is E. puercensis, so called from its occurrence in the Puerco beds of the American Eocene. It was a slender, long-limbed creature, smaller than Phenacodus, with a long and heavy tail as in that animal. Like Phenacodus it was semiplantigrade, and shows more likenesses to the Creodonta. The skull is only known by a part of the lower jaw with teeth, and by the teeth of the upper jaw. The vertebrae are not entirely preserved, but enough remain to show that the animal had a tail of 16 or 17 inches, which is a considerable length when compared to its height, about a foot at the rump. In the fore-limb the most noteworthy point is that the ulna has a convex posterior border as in the Creodonts, the same border in Phenacodus being concave. The humerus is slender, with less-marked tuberosities. The fifth digit seems to have been less reduced. The phalanges seem to have borne horny sheaths somewhat intermediate between hoofs and claws. The pelvis is described as being, as is also that of Phenacodus, rather like that of the Creodonta. The right hind-limb is known in all its details. It appears that the bones are not serial but interlocking; this, however, on the views with regard to the relations of these two forms of tarsus mentioned on p. [198], does not militate against regarding Euprotogonia as the ancestor of the genus Phenacodus. The third toe is the pre-eminent one, the animal thus being Perissodactyle. The lateral digits are larger than in Phenacodus, and the metatarsals and the phalanges are slightly curved, which is again a Creodont character as compared to the perfectly straight corresponding bones of Phenacodus. It seems evident that this animal is to be looked upon as a more ancient type than Phenacodus, even if not as its actual ancestor.

Another group of the Condylarthra contains the genus Pertipychus and some others. Periptychus has the full dentition

of forty-four teeth, the molars being of course bunodont, with the three chief tubercles most developed. The bones of the tarsus interlock and are not serial, as they are in many other members of the Condylarthra. The astragalus has a shorter neck than in Meniscotherium, for example. It has in this a likeness to the same bone in the Amblypoda, to the primitive members of which, such as Pantolambda, this animal bears much resemblance. "Astragali and many skeletal bones of Periptychus rhabdodon and Pantolambda bathmodon are almost indistinguishable," observes Mr. Matthew. The fore-feet of this genus are unknown, but it would seem that it was plantigrade from the evidence of the hind-feet. There are several species of the genus.

Possibly, but not at all certainly, the Mioclaenidae, with the genera Mioclaenus and Protoselene, are to be referred to this same order of primitive Ungulates. It is only necessary to mention them here, because they show very clearly the primitive form of dentition of these early Eocene mammals. The teeth are quite complete and unbroken by a diastema. The canines are but little pronounced. The molars are not strictly tritubercular, but have a prevailing trituberculy. The nature of the feet is not known. Since the genus Protoselene, as its name denotes, shows an indication of a commencing selenodonty, it has been suggested that this group is the stock whence the Artiodactyles have been derived.

In any case, whether the particular comparisons that have been made as to the relationship of various forms of Condylarthra are valid or not, it seems to be plain that this group represents the earliest Ungulate stock, but little differentiated from the contemporaneous Creodonts.

Sub-Order 2. AMBLYPODA.

This group of extinct mammals has the following principal characteristics:—

They are large, semiplantigrade Ungulates, of heavy build and apparently elephantine gait. The dentition is for the most part complete as in other ancient groups, and the canines are in the later forms big tusks. The back teeth are brachyodont and ridged (lophodont). Both radius and ulna in the fore-limb, and tibia and fibula in the hind-limb, are well developed. The bones

in the carpus are alternating in position. The toes are five in both feet, and are very short. There is a hint of commencing "perissodactylism" in the fore-feet at any rate. The brain is small and the hemispheres smooth.

The Amblypoda, or Amblydactyla, are so called on account of their short and stumpy feet and toes. They were held by Professor Cope to be on the direct line of ancestry of both Perissodactyles and Artiodactyles, a view which is on the whole not accepted at present.

Fig. 114.—Skull of Protolambda bathmodon. × ¾. e.a.m, External auditory meatus; m, mastoid; m.f, mastoid foramen. (After Osborn.)

As is the case with other groups, the Amblypoda commenced existence as a sub-order with relatively small forms such as Pantolambda, the most ancient type known, which is in many respects a transition between the later forms and other groups of mammals such as the Creodonta.[^[120]] The race culminated and ended in the giant Dinoceras and Coryphodon, and spread into the Old World. In spite of their smooth and diminutive brain, these mammals were able to hold their own and to multiply into many species and genera; in this they were perhaps aided by their formidable tusks and by the horns which many of them possessed. The teeth seem to imply an omnivorous diet, which was quite possibly an additional advantage in the struggle for existence. It does not seem to be necessary to divide off the Dinoceratidae into a sub-order equivalent to the Coryphodontidae as was done

by Professor Marsh; the numerous points in common possessed by the members of both families forbid their separation more widely than as families.

The earliest types of Amblypoda belong to the genus Pantolambda, of which the species P. bathmodon was about four feet in length. As restored it seems to have had proportionately short fore- and hind-limbs, and it had a long tail. It was apparently plantigrade, and would have had not a little likeness to a carnivorous type. The skull has no air cavities, such as are developed in the later types from the Lower Eocene, e.g. Coryphodon; Pantolambda is from the basal Eocene. The frontal bones show no trace of the horns that are developed in subsequent forms; the nasals are comparatively long; the zygomatic arch is slender. The molar teeth are in the primitive form of trituberculy, and the premolars, as is so often the case with primitive animals, are unlike the molars in form, being less markedly selenodont. As to the vertebral column, the dorsal vertebrae appear to have had short spines, which argues, as it does also in the case of the larger and heavier Coryphodon, a feebleness in the development of ligaments and muscles supporting and moving the head. The scapula seems to have the same peculiar leaf-like form that it has in the later Coryphodon.[^[121]] This primitive type shows an entepicondylar foramen in the humerus. It is interesting to observe that the posterior border of the ulna is convex, as in the Creodonts, and in the early Condylarthrous form Euprotogonia. In the subsequently-developed Amblypoda, as in the later Condylarthra, that bone acquires a concave outer border. In the carpus the os centrale is distinct. In the femur the third trochanter is well formed; it gradually dies out in later Amblypoda. The fibula articulates with the calcaneum. This species, according to Osborn, "typifies the hypothetical Protungulate, being more primitive than either Euprotogonia or Phenacodus."[^[122]]

Fig. 115.—Skeleton of Coryphodon radians. × 1⁄10. (After Osborn.)

The genus Coryphodon is known by a large number of species, of which the first was discovered in this country, and was represented merely by a jaw with some teeth. This was named by Sir R. Owen C. eocaenus, and was dredged up from the bottom of the sea off the Essex coast. A second specimen consisted of a single

canine tooth only, and was brought up from a depth of 160 feet during the making of a well at Camberwell. More abundant remains have since been found in North America.

This genus had a large head, and in some specimens traces of the "horn cores," so marked in the related Dinoceras, are to be noticed. The skull is broad behind and narrowed in front; the roofing bones show the cellular spaces so characteristic of the Elephant. The jugal bone, however, is not, as it is in the Elephant, placed in the middle of the somewhat massive zygomatic arch. As in some other primitive Ungulates (e.g. Phenacodus) there are twenty dorso-lumbar vertebrae, of which fifteen bore ribs.

The scapula seems to have possessed a peculiar leaf-like form, swelling in the middle and ending almost in a point above. It has a well-marked spine, and the acromion projects much. The fore- as well as the hind-feet are in a state of transition between plantigradism and digitigradism. It was at one time held that the animal was digitigrade as to the fore-feet and plantigrade as to the hind-feet. Though, as has been pointed out, it is a fact that the hind-feet are often on a different plane of evolution from the fore-feet, it seems that this amount of difference does not characterise any Ungulate, not excepting the genus now under consideration.

The toes are very spreading. The pelvic girdle is of great strength and broadness. The femur, as in the Perissodactyles, has a well-developed third trochanter; but whereas in this particular the hind-limb is Perissodactyle, it is Artiodactyle in the fact that the tibia and the fibula articulate with the astragalus and calcaneum. The ridged teeth have given the name to the genus.

A curious feature in the structure of the genus are the slender spines of the dorsal vertebrae, which contrast with the enormous ones of some other Ungulates—more curious in this genus, which is of heavy build, than in the lighter Pantolambda. The back of the animal is short, and the limbs are very spreading, so that the gait was doubtless shuffling. The large head, and short and heavy limbs and limb girdles added probably to its cumbrous walk or trot. The canines are great tusks, and spread out on both sides of the mouth.[^[123]]

The late Professor Cope, in 1874, described the probable appearance of the Coryphodon in the following words:—"The general appearance of the Coryphodons, as determined by the skeleton

probably resembled the bears more than any living animals, with the important exception that in their feet they were much like the elephant. To the general proportions of the bears must be added the tail of medium length. Whether they were covered with hair or not is of course uncertain. Of their nearest living allies, the elephants, some were hairy and others naked.... The movements of the Coryphodons doubtless resembled those of the elephant in its shuffling and ambling gait, and may have been even more awkward from the inflexibility of the ankle."

The most recent members of this sub-order come from the Middle Eocene beds, and are chiefly referable to the genus Dinoceras, with which Tinoceras and Uintatherium are at least very nearly related, if not identical. These creatures were of great size, larger than the earlier types which have been considered. They show a certain superficial resemblance to the Titanotheriidae, on account of the massive horn cores upon the skull. These horn cores are large upon the maxillae and the parietals, and are paired; on the nasals are smaller horns. The bones of the skull have air cavities. The incisors of the upper jaw are absent; the canines are enormous tusks, and the lower jaws are flanged downwards near the symphysis where these tusks border them. Contrary to what is found in the older types, where the position of the condyle of the lower jaw is normal, this prominence faces backwards in the Dinocerata. The same shortness of the spines of the dorsal vertebrae prevails in this group as in the other Amblypoda, though it is perhaps hardly so marked. The scapula has not the peculiar acuminate form that exists in Coryphodon, but is triangular and broad above. The limbs are elephantine, in that the angle between the humerus and the femur respectively, and the bones which follow, is not marked. The hind-limbs are especially straight. The tail is short as compared with that of the primitive Amblypoda. The Dinocerata are purely digitigrade. The entepicondylar foramen has, as in the Coryphodonts, disappeared. The os centrale of the carpus has become fused, and no longer exists as a separate bone. The fibula no longer articulates with the calcaneum, but both that bone and the ulna are well developed. The genus Astrapotherium is placed among the Amblypoda by some authorities.[^[124]]

Sub-Order 3. ANCYLOPODA.

The history of the discovery of the members of this order is very instructive as illustrating the dangers of laying too much classificatory importance upon detached fragments of animals. So long ago as 1825 terminal phalanges of a new creature were found in the Miocene of Eppelsheim, and sent to Cuvier. Cuvier named them "Pangolin gigantesque," deeming them, on account of their general form and cleft terminations, to pertain to the Edentata. In the same bed some seven years later were found certain teeth clearly of an Ungulate character, to which the generic name of Chalicotherium was applied. It was subsequently discovered that the teeth and the claws belonged to the same animal, and, later, further remains turned up which disclosed a creature having the anomalous composition of an Ungulate with decisively Ungulate teeth, but with the feet to a large extent like those of an unguiculate animal. The same confusion of characters occurs also, it will be remembered, in the distinctly Artiodactyle Agriochoerus (see p. [331]). Indeed the feet of the latter when first discovered were erroneously, as it now appears, referred to the present order of Ungulates under the name of Artionyx. It is probable that the genus Moropus of North America is a member of this group, and that it is probably congeneric with a somewhat different type of Ancylopod known as Macrotherium. It is also clear that Anisodon, Schizotherium, and Ancylotherium, if not congeneric with either of the two recognised genera, are at least very close to them.

Chalicotherium has a skull which recalls that of some of the earlier Ungulates; it has, however, no incisors at all, and no canines in the upper jaw; this feature has led to the belief that the animal is related to the Edentata, and that it is in fact a link between them and the Ungulata. The molars, like those of the Perissodactyla, are of the buno-selenodont type. It also agrees with that group (to which it has been approximated by several writers) in the. tridactyl manus and pes, and in the characters of the tarsus. But although tridactyl, the axis of the limb passes through the fourth digit. Chalicotherium is not mesaxonic, as are the Perissodactyles. Moreover, it has no third

trochanter, and the unguiculate claws have already been referred to. As to the latter, which are short, it is not the end phalanx but the first which is retracted; thus Chalicotherium differs markedly from both Carnivorous and Edentate types; for in the former it is the last phalanx which is retracted, while in the Edentates the same phalanx is flexed downwards. The limbs of Chalicotherium are nearly of the same size, and the animal seems to have been stout and quadrupedal.[^[125]]

Macrotherium, like the last genus, seems to have been common to both New and Old Worlds. It is to be distinguished by a number of characters. It is supposed to have been "semi-arboreal and fossorial"; the fore-limbs are much longer than the hind, the relative proportions of the radius and tibia being 70 to 29. The ulna was distinct from the radius, whereas in Chalicotherium the two are coalesced, or nearly so. Young specimens appear to possess a full set of incisors; whether this is the case or not with Chalicotherium is not known.[^[126]]

Homalodontotherium is sometimes placed in the group.

Sub-Order 4. TYPOTHERIA.

It is a little difficult to be confident that the Typotheria are rightly referred to the Ungulata, since they contradict two important Ungulate rules. They have clavicles, which are elsewhere missing, and the thumb looks as if it were opposable.[^[127]] An Ungulate is essentially a running animal, and has no need of a grasping finger. Still Typotheria are placed by most within the Ungulate series, though their undoubted likenesses to other groups, especially to the Rodentia, are admitted, and indeed emphasised. Cope places them definitely with the Toxodonts.

The Typotheria are an extinct group of smallish beasts, confined, like the Toxodontia, to South America, a region which during the Tertiary period, and into the Pleistocene, abounded with strange and varied types of Ungulate animals.

The earlier forms of Typotheria may be exemplified by some

account of the genus Protypotherium. This animal was of about the size of a Hyrax, which indeed it resembles in several points of structure. The teeth have the primitive number of forty-four, and they are close set, leaving no diastema; the molars are rootless and grow persistently; they are simple and Rodent-like in surface pattern. The shape of the lower jaw is like that of Hyrax, being rounded in outline posteriorly; there is no projecting angle as in the Rodents, and this remark applies to the Typotheria in general. The aspect of the Rodent lower jaw is characteristically different from that of Hyrax and the forms under consideration.

Some other characters of these early forms of Typotheria can be gathered from an inspection of other genera. In Icochilus both hand and foot were five-toed, and, as in ancient Ungulates generally, the bones of the wrist and of the ankle are serially and not alternately arranged. Moreover, an os centrale is present in the carpus. Both thumb and big toe were opposable. The skull has a remarkably Rodent-like appearance, but the palate is not so narrowed as in these animals.

In the more recent forms of Typotheria the dentition has become reduced. The canines are lost, and as the incisors are reduced also, to one on each side of the upper, and two on each side of the lower jaw, the likeness to a Rodent skull is increased. There is also evidence of a modification from the more primitive forms in the loss of one premolar or even more, in the alternating bones of the carpus, in the disappearance of the centrale, and in the loss of a toe upon the hind-foot. In these more recent forms the fibula articulates with the astragalus instead of with the calcaneum. Typotheria of these more recent forms may be illustrated by the typical genus Typotherium. It has the reduced dental formula I 1/2 C 0/0 Pm 2/1 M 3/3; the molars are simple in pattern, and much like those of Toxodon. The upper incisors are powerful and curved, but are surrounded by a layer of enamel, which is not limited to the anterior face, as it practically is in Rodents. The sacrum is composed of a large number of vertebrae—some seven—a state of affairs which recalls the Edentata. The shoulder blade is not Ungulate in form. It has a strong spine, with an acromion and a well-developed metacromion. The terminal phalanges are enlarged and hoof-like.

In the genus Pachyrucos there are three premolars, otherwise

the formula is the same as in Typotherium. The animal seems to have had nails rather than hoofs. The thumb was opposable. The fibula is fused below with the tibia, whereas in the last genus these two bones are quite separate from each other.

Sub-Order 5. TOXODONTIA

The group Toxodontia,[^[128]] like so many others, is exceedingly hard to define. Nor are its limits any easier to mark out than many others of the groups of Ungulates. It will be best perhaps to give an account of Toxodon, and of a few types which seem to lie near it in the system, and then to indicate how far they resemble or depart in structure from other Ungulates. Toxodon itself is known from complete skeletons. It lived in Argentina during the "Pampean" period, which seems to be of the Pleistocene age. A large number of species, however, have been described, some of which seem to go farther back in time, and to have existed during the Miocene period further south in Patagonia.

The size of this creature was about that of a large Rhinoceros; it has a bulky body and a large head, which was borne low down, on account of the bending downwards of the anterior vertebrae; in this aspect the figures of the skeletons recall Glyptodon and similar Edentates. The beast was discovered by Darwin, and originally described by Owen. "During his (Mr. Darwin's) sojourn in Banda Oriental," writes the Rev. H. Hutchinson, "having heard of some 'giants' bones' at a farmhouse on the Sarandis, a small stream entering the Rio Negro, he rode there, and purchased for the sum of eighteenpence the skull which has been described by Sir R. Owen. The people at the farm-house told Mr. Darwin that the remains were exposed by a flood having washed down part of a bank of earth. When found, the head was quite perfect, but the boys knocked the teeth out with stones, and then set up the head as a mark to throw at." The whole of the Pampean area is a valley of dry bones, and the remains of Toxodon are abundant there. The skull of Toxodon is not unlike that of a horse in general aspect; but the orbit is not separated from the temporal fossa. The premaxillae are furnished above with a slight protuberance directed towards

the free end of the nasals, which may be related to the presence of a short proboscis. The zygomatic arch is strong and broad: the mandibles are provided with a long symphysis. The dental formula is I 2/3 C (0-1)/1 Pm 4/(3-4) M 3/3. The teeth are prismatic and hypselodont, growing from persistent pulps. The molar teeth are slightly arched in form, whence the name of Toxodon, "bow teeth." The strong chisel-shaped incisors suggest the Rodents and Hyrax. The cheek teeth, moreover, are by no means unlike those of Rodents in their pattern. They are at any rate not at all like those of existing Ungulates. The small size of the canine and of the first premolar produces a diastema in the tooth series. The sacrum consists of five vertebrae, and the ischium does not articulate with it.

The shoulder blade has a strong spine, but only a rudimentary acromion; nor is the coracoid well developed. The radius crosses the ulna, as in the Elephant; the whole fore-limb is shorter than the hind-limb, which must have exaggerated the hang-dog expression of the creature when alive. The elements of the carpus interlock in the modern fashion. Those of the tarsus, however, are primitive in lying below each other without alternation. The carpus has a centrale. The fibula articulates with the calcaneum. The femur has no third trochanter. There are three toes to all the limbs. It is clear that this assemblage of characters will not allow the placing of Toxodon in any living Ungulate order. If the middle toes appear by their slight pre-eminence to approach the Perissodactyle form, the peculiar surface contour of the molar teeth, letting alone the absence of a third trochanter on the femur, will not permit this classification.

Allied to Toxodon is the genus Nesodon. It was so named from an "island lobe" on the inner side of the upper molars. This creature, smaller than Toxodon, also differs from it in the fact that the dentition is complete, and in the pattern of the molars, which is rather more complex. There is still the slight projection upon the premaxillary bones, but the nostril is directed more forwards than in Toxodon. The zygoma, too, is massive. The second pair of incisors in the upper jaw and the outer (third) pair in the lower jaw form biggish tusks in the adult. These and the molar teeth are, however, finally rooted, and do not grow, as in Toxodon, from persistent pulps. The genus is from the older Tertiary of Patagonia. Five or six species have been described. Some are as large as a Rhinoceros, others as small as a Sheep.

There is no doubt about the close alliance of the two genera just referred to. It is more doubtful whether Homalodontotherium and its allies should be placed, as they often are, in the neighbourhood of the Toxodonts. Homalodontotherium owes its name to its even row of teeth without a diastema. It was a creature of equally large size with Toxodon, and also came from the Tertiary strata of Patagonia. The teeth are the typical forty-four, and the molars like those of a Rhinoceros; they are, however, brachyodont and not hypselodont as in Toxodon. This genus, however, shows an important difference from the Rhinocerotidae and from the other Toxodontia in the fact that it was five-toed, and that the bones of the carpus and tarsus are set in relation to each other in the linear serial fashion.

Undoubtedly a near relative of Homalodontotherium is Astrapotherium. This creature was of equal bulk, and was also Patagonian in range. The teeth are reduced in number, but the animal was provided, like a Wild Boar, with great tusks, which were, however, formed by the incisors. This animal is very imperfectly known; it is the form of the molars and the large size of the incisors which have led to its association with the Toxodontia. As to the resemblance of the teeth of this genus and of Homalodontotherium to those of Rhinoceros, it is difficult to regard it as evidence of near affinity. The likeness is probably to be looked upon as a case of parallelism in development. Exactly the same explanation is possibly to be given to the likeness which the teeth of Toxodon and Nesodon show to Rodents, or even to Edentates. As to their affinities Zittel observes:—

"The entirety of their osteological characters argues for the Toxodon a separate position in the neighbourhood of the Perissodactyla, Proboscidea, Typotheria, and Hyracoidea. The relations to the Rodentia rest mainly upon the converging development of the teeth, not upon true relationship."

Sub-Order 6. PROBOSCIDEA.

Large vegetable-feeding animals, usually scantily covered with hair, and with the nostrils and upper lip drawn out into a long proboscis. Digits five on both limbs. Femur and humerus not bent upon lower leg and fore-arm in a position of rest. Skull

with abundant air cavities in the roofing and other bones. The incisors are developed into long tusks, which exist in the upper jaw alone, in the lower jaw alone, or in both jaws. There are no canines. The molars are lophodont. The clavicle is absent. The femur has no third trochanter. The bones of the carpus are serially arranged and do not interlock. The stomach is simple. The brain has much convoluted cerebral hemispheres, but the cerebellum is completely uncovered by them. The intestine is provided with a wide caecum. The testes are abdominal. The teats are pectoral in position. The placenta is non-deciduate and zonary. There are two venae cavae superiores.

The position of the limbs in the Elephant tribe is unique among living animals: their straightness that is to say, and the absence or very slight development of angulation at the joints of the limb bones. This same feature has been observed in the extinct Dinocerata and in the Titanotheria. It must not, however, be assumed from the resemblance to these ancient forms that there is much affinity between them and the Proboscidea, or that the latter have retained an ancient feature of organisation. The oldest Ungulates for the most part, and the Creodonts to which they are undoubtedly related, have much bent limbs. It must be considered, therefore, that the arrangement obtaining in the Elephants is purely secondary. Professor Osborn has put forward the reasonable view[^[129]] that the vertical limbs of all these colossal creatures are due to "an adaptation designed to transmit the increasing weight" of these animals. The huge bulk of the body is better borne by vertical pillars than by an angulated limb. Other points, however, such as the exposure of the cerebellum, the two venae cavae, the five digits, and the absence of a third trochanter, argue a low position for the Proboscidea in the Eutherian group.

The group can be readily divided into two families, the Elephantidae and the Dinotheriidae. We will commence with the former.

The Elephants proper, Elephantidae, differ from the Dinotheriidae in, and are characterised by, a number of anatomical features. They possess long tusks (incisors) either in both jaws, or, if only in one jaw, in the upper. The molar teeth are very large—so large that only a few of them are simultaneously in use. There are but three definable genera of Elephantidae, of which

Elephas alone survives. This genus also includes many extinct forms, both American and European, as well as Asiatic and African. The entirely extinct genera are Stegodon and Mastodon. The group is clearly one dwindling towards extermination. From the Middle Miocene downwards these great "pachyderms" have existed; and from the Miocene up to Pleistocene times they were almost world-wide in range and numerous in species.

The genus Elephas comprises usually large, but occasionally (the pygmy Elephant of Malta) quite small forms. The external features of the genus differ slightly in different species, and will therefore be described in relation to those species which we shall notice here. The vertebral formula is C 7, D 19-20, L 3-5, Sa 4-5, Ca 24-30, or even more.

The bodies of the vertebrae are remarkable for their shortness and for the very flattened articular surfaces.

Fig. 116.—A section of the cranium of a full-grown African Elephant, taken to the left of the middle line, and including the vomer (Vo) and the mesethmoid (ME); an, anterior, and pn, posterior narial aperture. × 1⁄12. (From Flower's Osteology.)

The skull is large and massive. Its large and heavy character is, as has been stated in the definition of the sub-order, due to the immense development of air cavities in the diploe; the diameter of the wall of the skull is actually greater than that of the cranial cavity. These cavities are not obvious in the young animal. They are most conspicuous in the roofing bones of the skull, but are seen elsewhere, and thicken the basis cranii,

the maxillae, and so forth. This state of affairs, together with the presence of the huge tusks, has, as it were, pushed back the nasal orifices to near the top of the skull in a very Whale-like fashion. As in the Cetacea, the nasal bones are limited in size, and the premaxillae send up processes to join the frontals and the nasals. There is a straight and somewhat slender zygomatic arch, but the orbit is not separated from the temporal fossa. The malar bone is small, and, as in Rodents, forms the middle part of the zygoma. This is not the case with most Ungulata. The symphysis of the mandibles forms a spout-like rim. The scapula has a narrow prescapular, but a very wide postscapular region. The spine has a strong process projecting backwards from near its middle; this is a point of likeness to certain Rodents. No Elephant has a clavicle. The most remarkable feature about the fore-limb is the separation and crossing of the radius and ulna. The arms of these animals are permanently fixed in the position of pronation. The foot is short, and the bones of the carpus are serially arranged. There are, however, traces of a commencing interlocking of these bones in many forms. The hind-feet are somewhat smaller than the fore-feet, and the tibia and fibula are both developed.

As to the teeth, this genus is to be distinguished from allied forms by the presence of tusks in the upper jaw only. These tusks have no bands of enamel such as characterise those of Mastodon. They are incisors. There is, however, a trace of the former enamelling in the shape of a patch at the tip, which soon wears away. The molar teeth of Elephas are so large that the jaws cannot accommodate more than at the most two and a part of a third at a time. These are gradually replaced by others to the number of three, the replacement of teeth suggesting that of the Manatee. Each molar is deeply ridged, the interstices between the ridges being filled up with cement. As the tooth wears away, therefore, the surface continues to be flat. Each ridge consists of a core of dentine surrounded by a coat of enamel. The number of these ridges varies greatly from species to species. The Indian Elephant is one of those which have the greatest number of plates in a single tooth, as many as twenty-seven.[^[130]] Of the six molars which

eventually appear, the first three are considered to correspond to premolars. But successional teeth are rare in the genus; that is to say as far as concerns the molars, for the tusks have their milk forerunners. As to the molars it is apparently only E. planifrons which certainly shows a milk dentition. In Mastodon and older types a milk dentition is commoner.

The viscera of the Elephant have been examined by many zoologists. The latest paper, dealing chiefly with the African species, but containing facts about its Indian congener also, is quoted below.[^[131]] The Elephant is remarkable in possessing, in addition to the three usual pairs of salivary glands present in mammals, a fourth, situated in the molar region, and opening on to the cheek by many pores. This gland is especially well developed in Rodents. There is a gland which may be mentioned in this connexion, though it opens externally between the eye and ear, known as the temporal gland; its use does not seem clear. The thoracic cavity of the Elephant, as may be inferred from the large number of ribs, is very large as compared with the abdominal.

The stomach is simple in form, and the epithelium of the oesophagus does not extend into it as is the case with the Horse and Rhinoceros. A gland or a collection of smaller glands occurs in the stomach, and recalls the "cardiac gland" of the Wombat and the Beaver, also that of the Giraffe. The large intestine is long, rather more than half the length of the small intestine. The caecum is well developed in these animals. The liver has a very simple form, being but slightly lobulated. It is actually only bilobed, but it is important to notice that this division does not correspond to the two halves of the liver. As shown by the attachment of the suspensory ligament, one half consists of the left lateral lobe alone, the other half embracing the remaining primary lobes. The simplicity of the liver looks like an archaic character. No Elephant has a gall-bladder. The lungs again are simple in form through their slight lobulation. Each half in fact is without subdivisions, and is of a triangular form. In this the Elephants resemble the Whales, as in the simple liver. In both cases probably the likeness is due to the permanence of primitive features of organisation. The brain[^[132]] of the Elephant

has hemispheres which are extremely well convoluted; but they leave the cerebellum entirely uncovered. This suggests a brain which is a great specialisation of a low type. The brain has been particularly compared with that of the Carnivora, with which group the Elephants agree in the characters of the placenta. It is, however, always a matter of the very greatest difficulty to compare the brains of mammals belonging to different orders.

There are but two living species of Elephant, of both of which we shall now proceed to give some account. Only a few of the rather numerous fossil forms can be touched upon here.

The African Elephant, E. africanus, has been sometimes referred to a distinct genus or sub-genus, Loxodon, by reason of the lozenge-shaped areas on the worn grinding-teeth. It lives, as its name denotes, in Africa. This species has a number of external features which enable it to be distinguished from the Oriental Elephant. The head slopes back more, and has not the two rounded bosses which give so wise a countenance to the Indian species. The ears are very much larger. The tip of the trunk has a slight triangular projection on both the lower and the upper part of the circumference of the aperture. There are four nails on the fore-feet and three on the hind. As in the Indian form, the toes are all bound together, and do not appear for any part as free digits. A thick pad of fat, etc., makes the animal when alive look as if plantigrade, whereas it is, as a matter of fact, digitigrade. In internal features the most prominent difference from E. indicus is in the molar teeth, which are ridged by much fewer ridges. The outside number for a single tooth in the present species is 10 or 11. In Elephas indicus on the other hand there are as many as 27.

The African Elephant, thinks Sir Samuel Baker, reaches a height of about 12 feet, and it will be remembered that the notorious "Jumbo" was found to be 11 feet high at the shoulder. The tusks are found in both sexes, as in the Indian beast, but are relatively larger in the female in the species now under consideration. It is also a rather more active creature, and is more savage;[^[133]] however it can be tamed, as is shown by several

specimens which have been and are in the possession of the Zoological Society, and other proprietors. It was apparently used in the past. Certain Carthaginian coins are stamped with a figure of the African Elephant; but in Africa no attempts are now made to utilise this creature except for food and ivory.

Fig. 117.—African Elephant. Elephas africanus. × 1⁄56. (After Sir Samuel Baker.)

The meaning of an Elephant as an emblem. upon a coin appears to be eternity, and there is no question but that the

Elephant is a long-lived animal. It is said that it hardly reaches proper maturity before forty, and that 150 years is not beyond probability in the way of longevity. Even longer periods have been assigned to it.

The tusks of the Elephant are by no means necessarily sexual adornments, used for fighting purposes only. The African Elephant is a most "industrious digger," and grubs up innumerable roots as food. It appears to be a fact that during these operations the right tusk is mainly used, and in consequence that tusk is shorter as well as thinner than the other. Two average tusks would weigh respectively 75 and 65 lbs., the latter of course being the weight of the more worn right tusk. These weights, it should be observed, by no means indicate the limits to which finely-developed tusks can attain. The very heaviest tusk known to Sir Samuel Baker[^[134]] weighed 188 lbs. This was sold at an ivory sale in London in the year 1874. The pace of the African Elephant, says the same authority, is at most at the rate of fifteen miles an hour at first, and of course in a furious rush. This pace cannot be kept up for more than two or three hundred yards, after which ten miles an hour is a better approximation to the rate which can be kept up for long distances.

The Indian Elephant, Elephas indicus (or Euelephas indicus, if the genus Loxodon is to be accepted), is better known and has been longer known than the African. It occurs in India and Ceylon, and in some of the Malayan islands, the Elephants of which latter parts of the world have been regarded as a distinct form, an apparently unnecessary procedure.

Fig. 118.—Indian Elephant. Elephas indicus. × 1⁄54. (After Sir Samuel Baker.)

This species does not stand so high at the shoulder as the African; its back is more rounded in the middle. The trunk has but one pointed tip; there are five nails on the fore- and four on the hind-feet. As this species comes from India and the East, it has been longer as well as better known than the African form. Thus many of the stories and legends that have congregated round Elephants apply really to this form. As is well known, the Indian Elephant is much used as a beast of burden, and for other purposes where its huge strength renders it invaluable. But its great drawback as a servant of man is its great independability. On the one hand we have furious, vicious, and generally unreliable

Elephants, and on the other perfectly docile creatures, who obey the slightest hint from their driver. Huge though the Elephant is, it is frequently a timid beast. Sir Samuel Baker relates how one which he was riding fairly bolted at the sight of a Hare. To

be bolted with by an Elephant is far from pleasing, though a rather exciting event. It makes for the nearest jungle at once, being, much more than the African species, an inhabitant of forest. And in rushing through the dense undergrowth, the occupiers of the Elephant's back are apt to be swept off or cut to pieces by innumerable thorns.

Elephants, no doubt of the Indian species, were used by the Persians in battle, and from fifteen which were captured at the battle of Arbela some notes were drawn up by Aristotle. In stating that the animal reaches an age of 200 years, the naturalist and philosopher was probably not very far out. The mode of Elephant-catching as related by Aristotle is that pursued at the present day. Then, as now, tame Elephants were made use of as decoys. Pliny,[^[135]] who was apt to confound fact and fiction in a somewhat inseparable tangle, had something to say about Elephants, both Indian and African. Serpents, he thought, were their chief enemies, which slew them by coiling round them and thrusting their heads into the trunk, and so stopping respiration. In Europe Elephants were first seen in the year B.C. 280. Pyrrhus used them in his invasion, and copying his example the Romans themselves learnt to use Elephants. The first Elephant seen in England arrived in the year 1257, presented by the King of France to Henry III. It was kept in the Tower (for long afterwards a menagerie), and died at twelve years of age. Much use of the Elephant has been made in symbols. We have spoken of the African Elephant on Carthaginian coins as an emblem of eternity. The Oriental Elephant resting on the back of a tortoise and supporting the world is the same idea; and it is instructive to note that remains have been found in the Siwalik Hills of a tortoise which would have been actually big enough to support the creature, even "Jumbo," who weighed 6½ tons. Another symbol is that of an Elephant upon whose back is a child with arrows; this occurs on a medal of the Emperor Philip. It can perhaps hardly signify the eternity of a strong human feeling!

The intelligence of the Elephant has been both exaggerated and minimised. Perhaps the most elaborate attempt to endow the beast with unusual mental perceptions is that of Aelian, who related that an Elephant carefully watching his keeper, wrote after him with his trunk letters upon a board. That the animal does

possess a good deal of brains, seems to be shown by the way in which a well-trained animal will obey the slightest sign of the mahout in India. According to Sir Samuel Baker, localities which produce in abundance particular kinds of fruit are remembered, as well as the time at which the fruit will be at its best. Stories of revenge, which are numerous enough, attest, so far as their data are to be accepted as accurate, the power of memory possessed by the Elephant.

In spite of their longevity, however, Elephants, unlike Rome, have not been built for eternity. We can only find two living species; but in past times Elephants were very numerous. They commenced, so far as we know, in the Miocene.

The existing forms are known in a fossil, or at least sub-fossil state, from diluvial deposits; and it is interesting to note that the African Elephant had formerly a wider range than now. Its bones (described as E. priscus) have been met with in Spain and Sicily.

One of the best known of completely-extinct Elephants is the Mammoth, E. primigenius. This great Elephant in most respects more nearly approached the existing Indian Elephant. The teeth have quite as numerous plates. The tusks were enormous, reaching a maximum length of 15 feet; they were much curved upwards as well as outwards. A large tusk weighs as much as 250 lbs. The Mammoth was of exceedingly wide range. Not only was it found in various parts of Europe, but it was especially abundant in Siberia, as is exemplified by the fact that for the last two hundred years as many or more than 100 pairs of tusks annually have been sold from that region. It also occurred in America together with forms at least not far removed from it, such as E. columbianus. Mammoths have been more than once found as entire carcases in the frozen soil of Siberia. The first was discovered in the year 1799, and rescued some years later for the St. Petersburg Museum. This example showed that the Mammoth, unlike existing Elephants, was covered with thick wool mingled with long and more bristly hairs of some 10 inches in length. The softer wool formed a kind of mane beneath the neck, which hung down as far as the knees. Another carcase was discovered later by Lieut. Benkendorf, who did not save it, but was nearly swept along with it into the sea by a flood. These creatures died in the position in which they were found by being bogged when in search of vegetation or water.

How primeval man, with his inferior weapons, slew the Mammoth is not easy to understand; but that they were contemporaneous is clearly shown by associated remains, and by the notorious sketch of the Mammoth on a piece of its own ivory, in which curved tusks and a forehead like that of an Indian Elephant are plainly to be seen. Although it was only so recently as the year 1799 that an example of this great creature was actually studied on the spot, and removed to St. Petersburg, the existence of Mammoths and of ivory is a matter of much more ancient knowledge. M. Trouessart relates[^[136]] that fossil ivory was known to the Greeks. Theophrastus spoke of ivory imbedded in the soil, and the tusks were recovered by the Chinese. It is a curious fact that the Chinese described and figured the Mammoth as a kind of gigantic Rat. The likeness between the elephantine molar and that of Rodents has been commented upon; but the existence of its tusks below the level of the ground led the Chinese Natural Historians to consider that the ways of life of the Mammoth were those of the Mole. As to the carcases themselves, the Chinese said that the flesh was cold, but very healthy to eat. This expression can hardly be explained, except upon the view that fresh carcases were known to that people long before they were known to us of the Western world. The value of the Mammoth ivory was known to antiquity; the famous Haroun-al-Raschid gave to King Charlemagne not only a pair of living Elephants, but a "horn of Licorne," which seems undoubtedly to have been a name for the tusks of the Mammoth. For in an account of the sacred treasures of Saint Denis, published in the year 1646, the author states this to be the fact.

The causes of the disappearance of the Mammoth are not easy to understand. Some held that it was a naked animal like the existing Elephants, and that the lowering of the temperature in Siberia proved fatal; it is, of course, now certain that it was clothed with dense woolly hair. Along with the bogged corpses of the great pachyderm, numerous trunks of pine-trees have been found, together with associated remains of other animals now extinct in that neighbourhood. Thus it is plain that Siberia was once covered by mighty forests, through which the Mammoth roamed. The decay of these forests, upon whose branches the Elephant fed, as is attested by the remains of pine leaves found

in the interstices of its teeth, was the signal for the disappearance of their most colossal inhabitant.

The large number of remains of this and of other extinct species of Elephas in this country gave rise to the supposition that they were Elephants brought over by Caesar to aid in the subjugation of these islands. The Rev. J. Coleridge (father of the poet) pointed out that though Caesar in his Commentaries made no mention of any such importation of Elephants, a passage in the Stratagems of Polyaenus expressly mentions that Cassivelaunus was confronted by the Romans with an Elephant clad in a coat of mail, by whose aid the crossing of the Thames was effected. At the time that attention was called to this (1757) it was not popular to hint at the possibility of fossils. So that fact, conveniently historical, served to explain away a difficulty. It is remarkable that the Elephant, common enough of course in Asiatic monuments, actually occurs in English architecture. Mr. Watkins, from whose interesting work (Natural History of the Ancients) a good many of the facts detailed here are drawn, tells us that the church of Ottery St. Mary has an Elephant's head sculptured on one of its pillars. The same ornament appears in Gosberton Church, Lincolnshire. Whether this has anything to do with a reminiscence of formerly existing Elephants is a hard question to answer. In this figure of an Elephant the trunk has a spiral representation, and the trunk of an Elephant is believed by some to be intended by the common "so-called Pictish ornamentation" in Scotland; this spiral alone is to be seen constantly. If it is a reduction of an Elephant to its simplest terms, it is highly interesting as an almost undoubted survival of remembrance of Elephants. For at such a period we cannot use the memories of Crusaders or others who may have visited the East to explain the facts. The sculptured Elephants' heads might conceivably be so explained.

The name Mammoth, thinks Mr. Watkins, may be derivable from the Arabic word Behemoth. He quotes a writer, who first described the beast in 1694, as using the two words indifferently. The Arabs, moreover, were then as they are now great ivory traders; and in the ninth and the two succeeding centuries explored the confines of Siberia, as they now do the forests of Africa, for ivory. The "Behemoth" of Job "eateth grass as an ox.... He moveth his tail like a cedar" (the Hippopotamus has a much more

stumpy appendage). "Behold, he drinketh up a river, and hasteth not" is surely much more suggestive of the copious draughts of an Elephant than the possibly equally copious but not so visible libations of a Hippopotamus.

The most ancient of the true Elephants (genus Elephas) is E. meridionalis. It is of the African type, i.e. the plates of the molar teeth are not abundant, and are not so many as in the existing E. africanus. It seems to have been one of the largest of Elephants, standing 4 metres high. Its remains are abundant in Europe, and are known also from England. Like this species E. antiquus is also of the African type. It was contemporary with man. Certain dwarf or "pony" races found in caves in Malta, and called Elephas melitensis or E. falconeri, are believed to belong to this species. Mr. Leith Adams, who described these[^[137]] remains, placed them in two dwarf species called by the names used above, and found associated with them a larger form, which he referred to E. antiquus. The existence of these animals in Malta seems to argue at least its former larger dimensions, and the presence of more abundant fresh water. The remarkable swimming capabilities of the Elephant do not necessarily imply either a former absence of land connexion or, on the other hand, its existence. Nor as a third possibility can it be suggested that the dwarf size argues an island of limited dimensions, when we bear in mind the huge tortoises of the Galapagos and some other islands. It is important to notice that Elephants of the African type (Loxodon) were not formerly absent from India. E. planifrons was one of these.

The genus Stegodon is so called from the fact that the molar teeth, seen in longitudinal section, present a series of roof-shaped folds, the interstices between which are not, or are, imperfectly filled up with the cement which in Elephas reduces the surface of the teeth to a level plane. This genus is exclusively Asiatic, and is Miocene to Pleistocene in time range. The number of ridges on the molars is small, not more than two. The incisors (tusks) have no enamel; the skeleton generally is like that of Elephas, between which and Mastodon the present genus is intermediate. Among the four or five species is S. ganesa (called after the Indian Elephant-headed divinity), with tusks 10 feet long, to be seen at the British Museum of Natural History.

The last genus of the family Elephantidae is Mastodon, so called from the structure of the molar teeth. These are provided with but few transverse ridges, not more than five, so that their structure is intermediate between those of Dinotherium and those of Stegodon. Between the ridges are sometimes isolated, boss-like protuberances (whence the name of Mastodon), produced by a subdivision of the ridges. There is either but little or no cement between the ridges. This genus differs from nearly all other Elephantidae by the possession of milk molars, which occasionally persist throughout life, the permanent dentition in those cases being a mixture of milk and permanent teeth, as has been (erroneously) stated of the Hedgehog.[^[138]]

The tusks (incisors) are sometimes present in both jaws, and as they have, during youth at any rate, a coating of enamel, the likeness to the chisel-shaped incisors of Rodents is patent. In connexion with the implantation of incisors in the lower jaw, many species have a prolongation of the bones of that part of the skeleton. In the bones, generally, there is not very much difference from Elephas, but the forehead is a little less pronounced. The genus existed from the Miocene and became extinct in the Pleistocene. It was nearly world-wide in range, being known from all four continents. Naturally with this very wide range was associated a large number of species. Zittel enumerates no less than thirty-two.

This genus is the only one of the Elephantidae which extended its range into South America, where the remains of two species occur. The bones of these great Elephants have attracted attention for some centuries. They were often held to be the bones of giants (as they actually were!), and in one case were ascribed to a deceased monarch, Teutobochus. The American Indians considered that equally gigantic men lived who were able to combat these great Proboscideans. There are legends of the Mastodons as living animals, which is quite probable, considering their geological age. There is a curious parallelism between the legends of two such widely-separated localities as North America and Greece. Buffon relates how among the Indians of Canada there was a belief that the Great Being destroyed both Mastodons and men of equal proportions, with thunderbolts. With this we may perhaps compare the story of the destruction of Typhoeus by Zeus, who

also used thunderbolts. One of the giants was not slain, but was compelled to stand and bear up the heavens. Atlas holds thus the position of the Elephant supporting the globe of Indian mythology.

Fig. 119.—Dinotherium giganteum. Side view of skull, 1⁄15th natural size. Miocene, Germany. (After Kaup.)

The genus Dinotherium, sole representative of the family Dinotheriidae, differs in a number of important particulars from the true Elephants. In the Elephants, if there is but a single pair of incisors, these are found in the upper jaw; in Dinotherium there is apparently but a single pair, but these are implanted in the lower jaw, the symphysis of which is much prolonged and greatly bent downwards, so that the tusks emerge at right angles to the long axis of the head, and are even bent backwards. The molar teeth are five in number on each side of each jaw and are bi- or tri-lophodont, not unlike those of the Tapir. There is no cement in the valleys between the ridges of these teeth, and there is a regular succession, the premolars being two and the molars three.[^[139]] All the teeth are in use at the same time,

their small size enabling them to be accommodated in the jaw together. The skull of Dinotherium is lower than that of Elephas or Mastodon. The bones of the skeleton generally are like those of Elephas.

Though a suggestion of marsupial bones attached to the pelvis has been discredited, there is no doubt that Dinotherium occupies the most primitive position among the Proboscidea; but at the same time it cannot be regarded as the ancestor of Elephants, as it is so much specialised in various ways. The incisors for one thing forbid this way of looking at the creature. It is an ancient genus found in beds of Miocene age in Europe and Asia. It is not known from America. The creature was larger than any Elephant. Eighteen feet in length has been assigned to it. The enormous weight of the lower jaw and tusks seems to argue that it was at least partially aquatic in habit, and that it may have used these tusks for grubbing up aquatic roots or for mooring itself to the bank. At first there were naturalists who considered it as an ally of the Manatee, and the skull is not unsuggestive of that of the Sirenia.

Pyrotherium has been referred to the Proboscidea; but our knowledge of that form is limited to a few teeth from Patagonian rocks of an uncertain age.[^[140]] They are simple bilophodont molars, very like those of Dinotherium. A tusk has been found in the neighbourhood of these teeth which may possibly belong to the same animal; but it is uncertain.

Sub-Order 7. HYRACOIDEA.

This group of small mammals contains only one well-marked genus which is usually named Hyrax, although Procavia seems to be the accurate term. Popularly these creatures are known as Coneys. They have a singular resemblance to Rodents, the short ears and much reduced tail, besides the squatting attitude adopted, contributing to this merely skin-deep likeness. They agree with other Ungulates in the structure of the molar teeth, which are much like those of Rhinoceros; in the absence of a clavicle; in the absence of an acromion; in the reduction of the digits of the limbs to four digits in the manus and three in the pes. On the

other hand they differ from most Ungulates in the incisors growing from persistent pulps, a point in which they resemble the Rodentia. The muffle also is split as in those animals. The Hyracoidea are peculiar in the fact that in addition to the caecum at the junction of the small and large intestines, there are a pair of caeca (bird-like in being paired) some way down the large intestine. The dorsal vertebrae are unusually numerous, 22. The adult dentition according to Woodward,[^[141]] who has recently examined the matter, is I 1/2 C (1/0) Pm 4/4 M 3/3, while the milk dentition is I 3/2 C 1/1 Pm 4/4.

Fig. 120.—Cape Hyrax. Hyrax capensis. × ⅛.

The inclusion of the canine of the permanent set of teeth in brackets signifies that it is the milk canine which occasionally persists. It should further be remarked about the teeth that they are both hypselodont and brachyodont, the extremes being connected by intermediate forms. Another peculiarity of the genus is the dorsal gland, which is covered with hair of a different colour to that covering the body generally. This is present in all species.

The genus Hyrax (the most recent authority on the subject, Mr. Oldfield Thomas,[^[142]] only allows one genus) is limited in its range to Ethiopian Africa and to Arabia, including Palestine, It does not reach Madagascar. Mr. Thomas allows fourteen species with two or three sub-species.

Some of the Coneys live in rocky ground, while others, formerly placed in the genus Dendrohyrax, frequent trees, in holes in which they sleep. The Coney of the Scriptures is familiar, who is "exceeding wise," though a "feeble folk." But the further observation that he "cheweth the cud but divided not the hoof," is obviously entirely wrong. As to the wisdom, it is said that this beast is too wary to be taken in traps; while the suggestion of chewing the cud is, according to Canon Tristram, to be interpreted in the light of a habit of working and moving its jaws which the animal has. The traveller Bruce kept one in captivity to see if it did really chew the cud, and found that it did!

CHAPTER X

UNGULATA (continued)—PERISSODACTYLA (ODD-TOED UNGULATES)—LITOPTERNA

Sub-Order 8. PERISSODACTYLA

Fig. 121.—Bones of the manus A, of Tapir (Tapirus indicus). × 1⁄5. B, of Rhinoceros (Rhinoceros sumatrensis). × 1⁄5. C, of Horse (Equus caballus). × ⅛. c, Cuneiform; l, lunar; m, magnum; p, pisiform; R, radius; s, scaphoid; td, trapezoid; tm, trapezium; u, unciform; U, ulna; II-V, second to fifth digits; V in B, and II and IV in C, represented by rudimentary metacarpals. (From Flower's Osteology.)

These Ungulates derive their name, which is that given by the late Sir Richard Owen, from the fact that the middle digit of the hand and foot is pre-eminent. As will be seen from Fig. 121, the axis of

the limb passes through the third finger, which is larger than any of the others, and is symmetrical in itself. In this the present group contrasts with the Artiodactyla, where the axis is not "mesaxonic," but where there are two digits, on either side of the axis, which are symmetrical with each other. This arrangement of the limbs is highly characteristic, but appears to be not quite universal. In the Titanotheres, which form a group of the Perissodactyles, the fore-limbs are not quite accurately mesaxonic. Nor on the other hand can all Ungulates which show the Perissodactyle condition be safely included in the present group. The ancient Condylarthra and the Litopterna show precisely the same state of affairs. But other features in their organisation lead to their separation from the Perissodactyles, of which, however, the Condylarthra are probably ancestors. The Litopterna on the other hand, which possess even one-toed members like Equus, are believed to represent a case of parallelism in development. The number of functional toes varies from four to one. In the ankle joint the astragalus either does not, or does only to a comparatively slight extent, articulate with the cuboid as well as with the navicular bone. Moreover the fibula when present does not as a rule articulate with the calcaneum. In the opposed group of Artiodactyles the precise reverse of these conditions obtains. It is usually stated as part of the definition of this group that they do not possess horns of the type of those met with in the Cervicornia and Cavicornia. But the strong bony bosses on the skull of many Titanotheres, so curiously reminiscent of those of the not nearly related Dinoceras and Protoceras, may well have supported horns of the Ox and Antelope pattern.

Fig. 122.—Bones of the manus of Camel (Camelus bactrianus). × ⅛. c, Cuneiform; l, lunar; m, magnum; R, radius; s, scaphoid; td, trapezoid; u, unciform. (From Flower's Osteology.)