CHAPTER XVII

HISTORY OF THE MARSUPIALIA

The marsupials are a group of more primitive structure and greater antiquity than any which we have yet considered, so primitive, indeed, that they are referred to a separate infraclass, the Didelphia or Metatheria. The order is one of very great variety in size, form, appearance, diet and habits, and mimics several of the higher orders in quite remarkable fashion. Herbivorous, insectivorous and carnivorous forms are all numerous, as well as arboreal, terrestrial, cursorial, leaping and burrowing genera. Some are like hoofed mammals in appearance and the Rodentia, Carnivora and Insectivora are also closely imitated in externals. With all this diversity, most unusual within the limits of a single order, there is such a unity of structure, that a division of the group into two or more orders is impracticable.

At the present time, marsupials are very largely confined to Australia and adjoining islands, where they constitute nearly the whole mammalian fauna, and it is in the Australian region that the remarkable diversity already mentioned is to be observed. There are found the phalangers, kangaroos, bandicoots, Tasmanian “devil” and “wolf,” and banded anteaters, not to mention many other curious creatures. In the western hemisphere only the opossums (_Didelphis_, _Chironectes_) and one very interesting relic of a long vanished assemblage, _Cænolestes_ of Ecuador and Colombia, are in existence to-day. The opossums, of which some twenty-three species are recognized, have their headquarters in South America, to which nearly all of the species are confined, North America having but two or three.

The more important American marsupials are given in the table below:

Suborder POLYPROTODONTA

I. DIDELPHIIDÆ. Opossums.

_Didelphis_, Opossum, Pleist. and Rec., N. and S. A. _Chironectes_, Water Opossum, Rec., S. A. _†Peratherium_, low. Eoc. to low. Oligo., N. A. _†Microbiotherium_, Santa Cruz. _†Eodidelphys_, do. _†Ideodidelphys_, Casa Mayor. _†Proteodidelphys_, †Cretaceous, S. A.

II. THYLACYNIDÆ. Predaceous Marsupials.

_†Cladosictis_, Santa Cruz. _†Amphiproviverra_, do. _†Prothylacynus_, do. _†Borhyæna_, do. _†Proborhyæna_, Deseado. _†Pharsophorus_, do. _†Procladosictis_, Casa Mayor. _†Pseudocladosictis_, do.

Suborder DIPROTODONTA

III. CÆNOLESTIDÆ.

_Cænolestes_, Rec., S. A. _†Zygolestes_, Paraná. _†Palæothentes_, Santa Cruz. _†Abderites_, do. _†Palæpanorthus_, Deseado.

IV. †GARZONIIDÆ.

_†Garzonia_, Santa Cruz. _†Halmarhiphus_, do. _†Cladoclinus_, do.

Suborder †ALLOTHERIA

V. †PLAGIAULACIDÆ.

_†Polymastodon_, up. Cretac. and Paleoc., N. A. _†Ptilodus_, do. _†Neoplagiaulax_, Paleoc., N. A.

VI. †POLYDOLOPIDÆ.

_†Propolymastodon_, Casa Mayor. _†Polydolops_, do. _†Amphidolops_, do.

Despite all their diversity of appearance and habits, the unity of structure among the marsupials is such that the formation of groups of higher than family rank is very difficult, and it is by no means certain that the suborders currently accepted correspond to the facts of actual relationship.

Except in certain extinct South American genera, there is very little change of teeth, only the last premolar in each jaw being replaced. Sometimes the temporary tooth is long retained in function and, more rarely, it is shed very early; while in several genera no replacement of teeth has been observed. There is a difference of opinion among naturalists as to the proper interpretation of the marsupial dentition. According to one view, all except the replaced premolar belong to the milk-series and the permanent series has been lost; the alternative and more probable belief is that the milk-dentition has been almost or completely suppressed. Whichever one of these interpretations be the right one, there is strong reason to maintain that the very limited amount of change is not a primitive condition, but a secondary one, for a series of rudimentary teeth is formed before those which are to become functional. The only reasonable explanation of such a condition is that it has been derived from one in which the normal succession and replacement of the teeth took place. Something of the same sort has been observed in the simplicidentate rodents. The marsupial dentition differs from the placental one in the usual number of four molars, instead of three, and frequently also in exceeding the normal total number of 44. The incisors are almost always of a different number in the upper and the lower jaw and are frequently more numerous than in the placentals.

The skeleton has several diagnostic characters, which are present throughout the order, though one or other of these features may be absent in particular instances. The skull has a very small brain-capacity and elongate face and jaws. In the placental mammals, the sutures between adjoining bones of the skull tend to close by coössification, and the separate bones are clearly distinguishable only in young animals; but in the marsupials the sutures remain open for a much longer period. The lachrymal is expanded on the face and the foramen is outside of the orbit. The tympanic is a mere ring and permanently separate from the other bones of the cranium, while a false bulla is formed by the inflation of part of the alisphenoid. In almost all marsupials there are large openings or vacuities in the bony palate. One of the most characteristic and constant features of the marsupial skull is in the conformation of the angle of the lower jaw, which is turned inward, or _inflected_, at nearly a right angle with the body of the jaw. It is true that one existing Australian genus has lost this character; and in some of the placental orders, especially the Rodentia, a somewhat similar structure may occasionally be found, but it is never quite the same as in the marsupials, in which it goes back to a remote antiquity.

There are very constantly 19 trunk-vertebræ, of which usually 13 are dorsals. The tail differs greatly in length in the various genera, but most of them have well-developed tails. An additional pair of elements, besides the three which are found in the placentals, enter into the composition of the hip-bones; these are the _marsupial bones_, slender, flattened rods, directed forward in the abdominal wall and diverging in V-shape. Save in a few genera, clavicles are present and of full size. The humerus may or may not have the epicondylar foramen, but the femur never has the third trochanter. The feet vary greatly in form and structure, in accordance with the habits, but there is a very widespread adaptation to an arboreal life, and even in terrestrial and burrowing forms more or less distinct traces of this arboreal adaptation may be noted. This fact has led to a generally accepted inference that all existing marsupials had an arboreal ancestry.

The soft parts and more especially the organs of reproduction are likewise very characteristic, and one or two of these peculiarities may be mentioned. (1) In the female, the vagina is double and on the abdomen is the pouch, or _marsupium_ (which gives its name to the order), a hair-lined bag, opening either forward or backward, which serves to carry the young and into which the teats open. A considerable number of species have lost the marsupium, while other species of the same genera retain it, and there can be little question that its absence is a secondary condition. (2) Except in one modern Australian genus, the marsupials have no true placenta, and the young are born in a very immature state, incapable of even swallowing. The new-born young are transferred to the nipples of the mother and are attached to these and fed by the pumping of milk into their mouths by muscular action of the mother. A special, though temporary, arrangement of the gullet and windpipe is provided, so that the helpless young animal shall not be suffocated by the entrance of milk into the lungs.

SUBORDER POLYPROTODONTA

This suborder, as is indicated by its name, is characterized by its numerous incisors, which are 5/4, or 4/3, and none of them is especially enlarged; by the large canines in both jaws, simple premolars and tritubercular upper molars. The members of this group are carnivorous or insectivorous in habit, and all the existing ones are of small or moderate size, though some very large extinct forms are known. Except in one Australian family, the feet are not “syndactyl,” a term which means the enclosure of two or more digits in one fold of skin. The only existing American representatives of the suborder are the opossums, the great majority of which are Neotropical in distribution.

1. _Didelphiidæ. Opossums_

In this family the dental formula is: _i_ 5/4, _c_ 1/1, _p_ 3/3, _m_ 4/4, × 2 = 50. The incisors are small and closely crowded together, the canines large and tusk-like, the premolars simple and of compressed-conical form; in existing species, the upper molars are triangular, each of the three main cusps is V-shaped and there are additional minute cusps along the outer border; the lower molars have a high anterior triangle of three pointed cusps and a low heel with several distinct cusps. The humerus has an epicondylar foramen and the feet are five-toed; in the manus all the digits are armed with claws and the thumb is but partially opposable, while in the pes the hallux is without a claw and completely opposable, making the foot much like that of a monkey.

The division of the existing opossums into genera has caused much difference of opinion and practice among naturalists; there are five groups, which by some are regarded as genera, and by others as subgenera, all modern members of the family being very much alike. The species _Didelphis marsupialis_, which is common in the eastern United States and extends through temperate North America, Central America and tropical South America, has a complete pouch and is chiefly arboreal and insectivorous in habit. In the woolly opossums (_Caluromys_) there is no pouch, and the young, when sufficiently advanced, are carried on the mother’s back, winding their tails around hers. In both of these genera the tail is long, naked and prehensile, but in the tiny species of _Peramys_ the tail is short and hairy. Another Neotropical genus, _Chironectes_, the Yapock or Water Opossum, is the only existing instance of an aquatic marsupial. It has light grey fur, striped with brown, and webbed hind feet; living chiefly in the water, it feeds upon crayfish, water-insects and small fish.

The derivation of the modern North American opossums is a matter of great uncertainty. The present distribution of the family, with by far the greater number of its species confined to the Neotropical region, is certainly suggestive of a South American origin, but such considerations are very untrustworthy guides in tracing the history of animal groups. No opossum has been found in any North American formation between the Pleistocene and the lower Oligocene, though in the case of such small animals, negative evidence must be accepted with caution. In the White River Oligocene many minute opossums have been found and referred to the European genus _†Peratherium_, though it so closely resembles the modern _Didelphis_ that many systematists do not make the distinction. In the Eocene, Paleocene and upper Cretaceous, opossums were represented doubtfully; the material is too incomplete for assured determination; in Europe they existed in the Oligocene and upper Eocene. In South America the family went back uninterruptedly to the oldest mammal-bearing beds of Patagonia, which may be upper Cretaceous. The opossums are thus the remnants of an exceedingly ancient group, whose beginnings are to be sought in the Mesozoic era and which was probably spread over all the continents. To all appearances, the whole group vanished completely from the northern hemisphere, but reëntered North America from the south at some time during the Pliocene or early Pleistocene and permanently established itself here.

The opossums are the most primitive of existing marsupials, especially the little South American genus, _Marmosa_, and are regarded, by some of the most competent students of the order, as closely representing the ancestral type of all the Recent families and genera, both of the Polyprotodonta and Diprotodonta.

2. _Thylacynidæ. Predaceous Marsupials_

By many naturalists this group of flesh-eating forms is included in the Dasyuridæ. The family never entered North America, but played a very important part in the Tertiary of South America. Three existing genera of the Australian region throw considerable light upon the South American types, and therefore some account of them will not be out of place here.

The largest of modern predaceous marsupials is the animal (_Thylacynus cynocephalus_) erroneously, but very naturally, called the “Tasmanian Wolf,” now confined to Tasmania, but occurring also in the Pleistocene of Australia. As “wolf” applied to a marsupial is misleading, it will be less confusing to employ the anglicized form of the generic name “Thylacine.” This animal is of the size of the small Prairie Wolf or Coyote (_Canis latrans_) and has very wolf-like appearance and habits. The muzzle is long and pointed, the ears erect and rather small, the tail long, very thick at the base and tapering to the end, not bushy, but covered with short, close-set hairs; the colour is greyish brown, with dark, transverse stripes on the posterior half of the back and base of the tail. Apparently the creature is in process of losing its stripes and acquiring the solid body-colour. The dental formula is: _i_ 4/3, _c_ 1/1, _p_ 2/2, _m_ 4/4, × 2 = 46; the incisors are small, the canines large fangs, and the premolars simple; the upper molars are tritubercular, with large inner cusp and postero-external cutting ridge, and the lower molars are trenchant, with low heel. The whole dentition is remarkably like that of many Eocene †creodonts, such as _†Sinopa_ and _†Tritemnodon_ (see p. 566). The milk-premolar is small and functionless and is shed very early. The skull is very wolf-like in appearance, but thoroughly marsupial in structure, and has the large palatal vacuities common in the order. The marsupial bones do not ossify and are evidently on the point of disappearance. There are five digits in the manus, four in the pes, the hallux being completely suppressed. In habits, the Thylacine is carnivorous and so destructive to sheep that the farmers have nearly exterminated it.

[Illustration: FIG. 296.—Thylacine, or “Tasmanian Wolf” (_Thylacynus cynocephalus_).—By permission of W. S. Berridge, London.]

The other forms to be mentioned belong to the closely allied family of the Dasyuridæ. The “Tasmanian Devil” (_Sarcophilus ursinus_) is now, like the Thylacine, confined to Tasmania, but remains of it have been found in the Australian Pleistocene; it has one less premolar in each jaw, giving the formula: _i_ 4/3, _c_ 1/1, _p_ 2/2, _m_ 4/4, × 2 = 42; there is no milk-tooth. The premolars are closely crowded and the molars resemble those of the Thylacine in construction, but are broader and heavier. The skull is disproportionately large, with shorter and wider muzzle and jaws than in the Thylacine; the tail is of only moderate length and somewhat shaggy; the hallux is wanting. In size and build, the Tasmanian Devil resembles a badger and has long and heavy fossorial claws on the fore feet; the hair is rough and shaggy, black in colour with white patches. The animal has received its name from its fierce and savage disposition and is almost as destructive to sheep as the Thylacine.

The five species of _Dasyurus_ are distributed through Tasmania, Australia and New Guinea and are called “Native Cats”; they are much smaller animals than the two preceding genera, not exceeding a domestic cat in size. As the Thylacine imitates a wolf and the Tasmanian Devil a badger, the dasyures resemble the civets. In them the dental formula is the same as in _Sarcophilus_, but the teeth have higher and sharper cusps. The head has a narrow, tapering muzzle and narrow ears; the body is long and the tail of moderate length. The limbs are short and slender and a small hallux is present in some of the species. The fur is grey or brown, with numerous white spots, and the tail is covered with long hair, but not bushy. The dasyures are largely arboreal and prey upon small mammals, birds and eggs.

Until the arrival of the true Carnivora from the north, their rôle was taken in South America by predaceous marsupials, which persisted as late as the presumably Pliocene beds of Monte Hermoso. Little is known of them in that stage, however, or in the older Paraná, but abundant material representing those of the Santa Cruz has been collected. Among these there was a considerable range of size and some variety of structure, and they all differed in certain respects from the modern Australian genera, differences which have led some authorities to deny the marsupial character of all these South American forms. The differences are of three kinds: (1) there are no vacuities in the bony palate; (2) the milk-dentition is less reduced, the canines and one or two premolars being changed; (3) the enamel of the teeth, in the only genus (_†Borhyæna_) which has been examined microscopically, resembles in its minute features that of the placentals and lacks the marsupial characters. Though by no means unimportant, these differences are altogether outweighed by the thoroughly marsupial nature of all other parts of the skeleton, and I cannot but agree with Dr. Sinclair[20] in including them in the same family with the Tasmanian Thylacine.

The genus _†Prothylacynus_ was especially like the latter and must have had a very similar appearance, though in the restoration (Fig. 297) the colour-pattern is changed to one of longitudinal stripes, as more probably pertaining to so ancient and primitive a form. The humerus had the epicondylar foramen, and a large vestige of the hallux was retained, though it could not have been visible in the living animal.

A more specialized Santa Cruz genus was _†Borhyæna_ (Fig. 244, p. 494), an animal of about the same length and height as _†Prothylacynus_ and the Thylacine, but much more massive and powerful. The skull was remarkable for the small size of the brain-case and the great spread of the zygomatic arches, which gave a rounded and almost cat-like appearance to the head, as is shown in the restoration (Fig. 244). In this genus the upper incisors were reduced to three, a very unusual thing among the Polyprotodonta, and the humerus had lost the epicondylar foramen. _†Prothylacynus_ and _†Borhyæna_ were the largest of the Santa Cruz flesh-eaters and no doubt pursued the smaller and more defenceless ungulates, but were hardly sufficiently powerful to attack successfully the larger hoofed animals, which were probably well able to defend themselves.

[Illustration: FIG. 297.—Santa Cruz predaceous marsupial (_†Prothylacynus patagonicus_) and †typothere (_†Interatherium robustum_). Restored by C. Knight from skeletons in the museum of Princeton University.]

[Illustration: FIG. 298.—Skull of _†Borhyæna_, Santa Cruz. (After Sinclair, Reports Princeton University Expeditions to Patagonia, Vol. IV.)]

[Illustration: FIG. 299.—Skull of small predaceous marsupial (_†Amphiproviverra manzaniana_), showing the punctured wound from a bite. Princeton University Museum.]

Associated with these larger predaceous marsupials were several much smaller kinds, ranging in size from a fox to a weasel, which must have preyed upon the abundant rodents and other small mammals and birds. One of these (_†Amphiproviverra_) had an opposable hallux, somewhat as in the opossums, and was therefore probably arboreal. An interesting specimen in the museum of Princeton University illustrates the pugnacity of these small creatures; it is a skull in which the left upper canine was completely torn out, the circular puncture of the enemy’s bite being unmistakable and the healed edges of the wound proving that the loss of the tooth was suffered during life. In structure, these smaller animals differed so little from the larger ones, that no particular description of them is needed. In the restoration of _†Cladosictis_ (Fig. 300) the spotted pattern of the Australian dasyures, or native cats, has been taken as a model.

In the Deseado formation the predaceous marsupials have been less abundantly found than in the Santa Cruz and there can be little doubt that the group is very inadequately represented by the material so far collected. Only two genera, known from lower jaws, have been described, but one of these (_†Proborhyæna_) is of interest because of its enormous size, far surpassing any of the Santa Cruz forms and equalling the largest modern bears. This is another illustration of the unusual relationship between the Deseado and Santa Cruz faunas, the older stage so frequently having the larger animals.

Predaceous marsupials of small size may be traced back to the Casa Mayor formation, but very little is yet known of them. There is no obvious difficulty in the way of their derivation from opossum-like forms, such as are found in the Cretaceous of North America and probably of South America also.

The relation of the South American to the Australian marsupials offers problems of unusual interest, a discussion of which would be impracticable here. Several alternative solutions of the problem have been offered and great differences of opinion exist with regard to it. To my mind the most probable suggestion is that a land-connection, by way of the Antarctic continent, existed in early Tertiary times, by means of which the ancestors of the Australian marsupials migrated, from South America, though this explanation is rejected by several eminent authorities.

[Illustration: FIG. 300.—Small predaceous marsupial (_†Cladosictis lustratus_) and rabbit-like †typothere (_†Pachyrukhos moyani_), Santa Cruz stage. Restored by C. Knight from skeletons in Princeton University and the American Museum of Natural History.]

SUBORDER DIPROTODONTA

North America never had any representatives of this suborder, but South America possessed many of them in the Santa Cruz Miocene and one genus (_Cænolestes_) has survived to the present time. Australia, on the other hand, has three well-defined families of the suborder, the kangaroos, phalangers and wombats, but no member of any of these has ever been found outside of the Australian region. So far as we know, therefore, the suborder is and always has been confined to the southern hemisphere.

[Illustration: FIG. 301.—Skull of _Cænolestes obscurus_, enlarged. (After Sinclair.)]

The modern South American genus _Cænolestes_ is a small, rat-like animal and very rare; it has been found only in Ecuador and Colombia. Its dentition is not at all typically diprotodont, but rather intermediate in character between the latter and the Polyprotodonta. The dental formula is: _i_ 4/3, _c_ 1/1, _p_ 3/3, _m_ 4/4, × 2 = 46. The upper incisors are small and of subequal size, though the second is somewhat the largest of the series, and the canine is considerably larger and more prominent than any of them. The foremost lower incisor is long and pointed and directed almost straight forward; the other lower incisors and the canine are minute and can have little or no functional value. The premolars are small and simple and the upper molars quadritubercular, the third one triangular, and the fourth very small and apparently about to disappear. Such teeth would seem to indicate a vegetable diet, but it is reported that the animal subsists chiefly upon small birds and their eggs. The skull, which is typically marsupial in all its characters, is most like that of the smaller Australian native cats (Dasyuridæ) and the feet show no signs of the syndactyly which all the other diprotodonts display so clearly. Dr. Gregory is “inclined to regard _Cænolestes_ and its allies as an independent suborder, an offshoot of primitive Polyprotodonts which has paralleled the Diprotodonts in certain characters of the dentition.”[21]

[Illustration: FIG. 302.—Lower jaws of Santa Cruz cænolestids, enlarged. _A_, _†Garzonia patagonica_. _B_, _†Abderites crassignathus_. _C_, _†Callomenus ligatus_. (After Sinclair, in Reports Princeton University Expeditions to Patagonia, Vol. IV.)]

Evidently, the animals of this series were extremely rare or absent in the areas where the known South American deposits of the Pleistocene and Pliocene were laid down, for there is a very long hiatus in their history from the Recent to the Santa Cruz, during which none has yet been found, except one genus (_†Zygolestes_) in the Paraná. In the Santa Cruz, however, there was a great abundance of these little marsupials, to which various generic names have been given and which displayed considerable variety in the forms of the teeth. Some (e.g. _†Garzonia_) agreed with _Cænolestes_ in having no trenchant shearing teeth; behind the large, procumbent lower incisor, followed four or five very minute teeth, which must have been nearly or quite functionless, succeeded by the well-developed molars. Other genera (_e.g._ _†Abderites_) had a similar dentition, with the important exception that the last upper premolar and first lower molar were enlarged and trenchant, together forming a shearing pair; these teeth were vertically fluted or ribbed in very characteristic fashion. The Australian phalangers have very similar trenchant and fluted teeth, but in that family the lower one of the pair is the last premolar, not the first molar. Marsupials of this type have not been found in formations older than the Deseado.

The relationship of these South American genera to the Australian phalangers is a very interesting question from the standpoint of mammalian distribution, but is not likely to receive a positive answer until something is learned regarding the history of the Australian family.

SUBORDER †ALLOTHERIA

[Illustration: FIG. 303.—Skull of Paleocene †allothere (_†Ptilodus gracilis_), enlarged, Fort Union stage. (After Gidley.)]

This extinct suborder is still very imperfectly understood, for it is known almost exclusively from jaws and teeth; so far, the skull of one genus and most of that of another have been obtained, but hardly anything of the skeleton. The †Allotheria were small or minute marsupials, herbivorous or omnivorous, which had lost all trace of the canines and had one pair of incisors above and below, which grew from persistent pulps and had a scalpriform, rodent-like character. The molars were composed of numerous tubercles (whence the name “†Multituberculata,” often applied to the group) arranged in two or three longitudinal rows, and the premolars were either like the molars, but of simpler pattern, or compressed, sharp-edged and trenchant. The †Allotheria were among the most ancient of mammals and have been found in the Triassic of Europe, the Jurassic of Europe and South Africa, the Jurassic and Cretaceous of North America and the Paleocene of both northern continents, while the South American Eocene (Casa Mayor) had certain problematical genera (†Polydolopidæ), which may be referable to the †Allotheria or to the _Cænolestes_ series. The suborder was thus preëminently a Mesozoic one and, with the doubtful exception of South America, it is not known to have passed beyond the limits of the Paleocene. There is not the least likelihood that any existing mammals were derived from the †Allotheria.

[Illustration: FIG. 304.—Head of _†Ptilodus gracilis_, about natural size. Restored from a skull in the United States National Museum.]

While the †Allotheria have an antiquity at least equal to that of any other mammals known, there were other groups in the Jurassic and Cretaceous, which, so far as may be judged from teeth alone, would seem to have been ancestral to the other marsupials and to the placentals. It would serve no useful purpose to describe these minute creatures, which are so very incompletely known, though to the specialist they are of the highest interest. The genera found in the Triassic of North Carolina may or may not represent the primitive mammalian stock.

* * * * *

The question of the origin of the Mammalia is still involved in great obscurity, and the most divergent opinions are held concerning it. It remains an unsolved problem whether the mammals were all descended from a common stock, or have been derived from two independent lines of ancestry, or, in technical phrase, whether the class is monophyletic or diphyletic. Assuming, as seems most probable from present knowledge, that the mammals are monophyletic, the question next arises: From what lower vertebrates are they descended? A great controversial literature has grown up around this problem, one party regarding the Amphibia and the other the Reptilia as the parent group. The palæontological evidence, while not conclusive, is decidedly in favour of the latter view. In the Triassic of South Africa is found a group of reptiles which approximated the mammals very much more closely than do any other known representatives of the lower vertebrates. While it is not believed that any of these Triassic reptiles were directly ancestral to the mammals, they did, to a very great extent, bridge the gap between the two classes and show us what the reptilian ancestors of the mammals were probably like.

With perhaps the exception of certain Insectivora, the Paleocene faunas contained few, if any, ancestors of modern mammals. These originated in some region which has not been identified, but may be plausibly conjectured to be central Asia, whence they migrated westward to Europe and eastward to North America, reaching both of those continents in the lower Eocene. From that time onward they increased and multiplied, becoming more and more differentiated through divergent evolution, until the existing state of things was attained. From the lower Eocene we are on firm ground, and, though very much remains to be learned, much has already been accomplished in the way of tracing the history and development of many mammalian orders. It has been my endeavour in the body of this book to sketch the better established and more significant parts of this marvellous story.