It thus appears that the general geological principle with which we started, of the more complete destruction by denudation and metamorphism of the earlier as compared with the later records of life upon the earth, receives ample support in the apparently sudden appearance of whole groups of complex and specialized forms in some of the earliest rocks; while the general imperfection of the geological record is made manifest by such facts as the very few and isolated remains of Mammalia in the Mesozoic rocks, although we know they must have existed in abundance throughout the whole Secondary and much of the Palæozoic periods. The great lesson we have to learn from the facts of paleontology is, that its negative evidence is at the best of but little value; but when this negative evidence is opposed to general principles established upon a wide basis of physical and biological research, it becomes absolutely worthless. Just in proportion as the series of fossiliferous deposits is more complete, and the fossil remains more abundant and varied, does the evidence for evolution and progressive development become more powerful. The difficulties are almost wholly dependent on incomplete knowledge, and on the assumption (which we have endeavoured to show is entirely unfounded) that the earliest traces of the fossil remains of any animal type which have been discovered can give, even approximately, the period of its first appearance upon the earth,

We find, then, that just as a study of the distribution of animals in space enables us to learn much of the immediately preceding condition of the earth's surface, and especially of the recent changes of land and water,- -so a study of the distribution of animals in time, when aided by the modern theory of evolution, gives us some knowledge of the physical condition and life of the earth in times beyond the reach even of geological history. (A. B. W.)

DISTRIBUTION OF VEGETABLE LIFE.

The literature in which the immense multitude of distinct kinds of plants which are dispersed over the earth's surface and form its vegetation has so far been described has necessarily been adapted to the divisions of political geography. The causes which have brought about the formation of such divisions have rarely, however, had anything in common with those which have determined the characteristic features, whether superficial or profound, of the floras of different countries. The great mass of catalogues and descriptive enumerations of the plants of such countries, the boundaries of which are for the most part quite artificial, are therefore ill adapted for bringing out any general conclusions as to the mode in which plants are distributed. It is only by making some kind of analysis of the often heterogeneous contents of such catalogues, and piecing together the results obtained from different sources, that any clue can be obtained to the approximate lines of demarcation of floras which are really naturally limited and characterized. The process is, however, enormously laborious, and, even apart from that, must for a long time to come be exceedingly imperfect in its application, owing to the immense tracts of land-and those with the most varied and copious vegetation-of the natural products of which our knowledge is still most defective.

Numerous attempts have, however, been made, notwithstanding the difficulty of the task, to map out the earth's surface into "regions of vegetation." The real significance of these regions will, of course, entirely depend upon the principles which have been relied upon in forming them. And in this respect the progress of geographical botany has been exactly similar to that of classification. The characteristic distinctions which were first seized upon in

either case proved or closer scrutiny to be superficial, and to bring about merely artificial and arbitrary assemblages. The doctrine of evolution has in fact effected the same revolution in both; it has shown in the one that community of descent is the real meaning of a natural classification, it has shown in the other that community of origin is the real key to geographical distribution.

Most of the writers on geographical botany have been content to set aside all considerations of origin and history in attempting to define the limits of botanical regions. They have not attempted to see in the peculiar features which such regions may possess anything more than adaptations to physical conditions working on plants created in great measure where they are found. Although, therefore, the literature of geographical botany has been useful in enabling the reader to realize the local features—the colouring, if one may so express it-of particular countries, the facts have hitherto been presented in a form void of any true significance. And these remarks apply to the system of Schouw (1833), which has been much employed, partially to that of De Candolle, and conspicuously to that more recently published by Grisebach. It is to the writings of Darwin, Hooker, Asa Gray, and Bentham that we must look for a real insight into the origin and dispersion of floras, and for the real causes of the existing distribution of plant life.

The first attempt to review the whole subject of plantdistribution from the modern point of view afforded by evolution is due to Bentham, who made it the subject of a presidential address delivered to the Linnean Society in 1869. Bentham's conclusions are based upon the experience of a long life devoted to systematic botany, and will probably always hold a fundamental position in the study of the subject; at any rate for some time to come, until the distribution of a large number of subordinate groups has been carefully worked out, the main points established by him are not likely to be materially modified.

The general facies of vegetation is obviously largely affected by purely physical causes. In the polar regions, arboreal and even shrubby plants become incapable of existence, and only small perennials which are safely covered up by snow during the long winter are able in the brief summer to expand their flowers and ripen their seeds. Putting aside for the moment the severances effected by large bodies of water and mountain chains, it is easy to see that the vegetation of the earth must have always been separable into three great latitudinal zones, two belonging to the north and south hemispheres respectively, and one dividing them lying between the tropics. The constitutents of the vegetation of these zones must always have had a certain homogeneity; very considerable divergences, however, have grown up within the zones themselves, owing to circumstances of geographical isolation. Even without these, distance alone, independently of isolation, would in time be sufficient to effect it. It is also obvious that the precise northern and southern limitations of such hypothetical zones must have varied with secular changes in the earth's climate, and when these changes have taken place over a broken configuration of land and sea, the intermixture of diverse floras must necessarily have become very complicated.

Underlying, however, the tangled fabric of the earth's existing floral covering, we may agree with Bentham1 in recognizing the existence of three tolerably ancient florasthe Northern, the Tropical, and the Southern.

I. The Northern is characterized by its needle-leaved Coniferæ, its catkin-bearing Amentace and other forest

'Presidental address to Linnean Society, 1869, p. 18.

trees deciduous in winter. and its vast assemblage of herbaceous types, Ranunculacea, Cruciferæ, &c. These spread over Europe, northern and central Asia, and great part of North America.

II. The Southern is broken up into numerous divergent floras. Their original connection is now traceable only in the common possession by two or more of them of large characteristic groups, such as Restiacea, Proteacea, Diosmeæ, &c., the subordinate divisions of which have been locally specialized. To this belong the floras of extratropical South America, South Africa, and Australia with New Zealand, to which must probably be added an area borrowed from the northern hemisphere in Mexico and California.

III. The Tropical is characterized by the predominance of mostly evergreen arborescent Polypetala (Anonaceœ, Meliaceae, Leguminosa, &c.), and gigantic Monocotyledons,. of which Palms, Scitamine, and Bambuse amongst grasses are especially striking.

1

I. THE NORTHERN FLORA.-This has been long divided into that of Old and New World by the severance of North America from Northern Asia, and by the barrier to an interchange of vegetation in the upheaval of the Rocky Mountain range. Nevertheless its marked continuity (with only a gradual east and west change in the arctic regions, but an increased divergency southwards) requires it to be treated as a whole. The Old and New World divisions of this flora, which, no doubt, began to diverge from the mere influence of distance, have now had that divergence immensely increased by isolation. According to Lesquereux, the essential types of the present arborescent flora of North America are indicated in the Cretaceous rocks of that country, and become more distinct and numerous in the Tertiary; and he believes that the origin of the existing American flora is American. The analogy between the Miocene flora of Central Europe and the present North American flora is unquestioned, and is greater than between the same fossil flora and that now existing in Europe. Lesquereux's conclusion is that the American element in the vegetation of Miocene Europe was derivative, and this is one of many illustrations of the curious observation of Asa Gray that plants have in general a greater tendency to migrate from east to west than from west to east. This Miocene flora was, however, gradually driven back again, and it is only as we travel from Europe to the East that we gradually find its traces getting stronger and stronger. Thus, as Oliver has pointed out, in passing from the Mediterranean to the Levant, the Caucasus, and Persia, we meet with living representatives of the Miocene genera Chamaerops, Platanus, Liquidambar, Pterocarya, Juglans, &c. Along the Himalayas and through China we trace other Miocene genera, Japan forming part of the same botanical region as Eastern Asia. Among the remarkable existing North American types which may be mentioned as reappearing in the Himalayas and Japan are Aralia quinquefolia. Phryma leptostachya, and Trillium erectum. One of the most interesting additional facts which has recently come to light is the occurrence of a species of tulip tree (Liriodendron) in Central China, which genus, though a member of the European Miocene flora, has in recent times been regarded exclusively characteristic of America.3 With respect to other American genera which are not necessarily part of the Miocene flora, the same general principle holds good. Bentham remarks, that while some, like Astragalus, have multiplied largely in both continents, "other genera, like Eupatorium, Aster, Phlox, Solanum, &c., very nume1 Geological Survey of Montana, 1871, p. 314.

Nat. Hist. Rev. 1862.
Moore, Journ. of Bot. 1875, p. 225.

rously represented in America, have transmitted or produced a smaller number in Eastern Asia, gradually diminishing westward till they disappear altogether or attain Western Europe in single species but little altered from American ones." "24 The Europeo-Asiatic genera, on the other hand, such as Crucifera, Umbelliferæ, &c., which are so dominant a feature in the existing Old World Northern flora, appear "to have left but few representatives in America, and those much more modified than the American races left in Asia.” Besides the internal migrations of the various constituents of the great Northern flora, its boundaries have been changed longitudinally under the influence of secular variations of climate alluded to above. The nature of these cannot be better summed up than in the words of Bentham :—5

"Where the chief portion of this great northern flora originated, and whether it may be best termed Scandinavian, or North Asiatic, little or no data; but, as observed by Hooker, it is probably one of or Caucasian, is a question for the determining of which we have the most ancient and widest spread, having at different epochs travelled over a great part of the globe. Shown by the researches of Lesquereux, as well as by the recent ones of Heer and others, to have extended far north during the warmer preglacial times, it must have been slowly driven southwards as the glacial epoch came on, and either then, or at some one or more other periods, have been for a time continuous, in two lines at least, into the southern heraisphere; for it has left traces still discernible, especially in its down at least to the Indian peninsula, and westward to the Abys herbaceous and mountain forms, in the mountains of tropical Asia, sinian and Cameroons mountains of Africa, and, again, down the Andes to the extreme south of America, where it is still luxuriant, and in a less degree in New Zealand, Tasmania, and Victoria. In all these migrations, whilst retaining a general identity, the flora must have undergone continual changes, losing species or other races of limited areas and propagation as their habitations became unfit for them, and gradually forming new ones when favoured by long-continued isolation or other requisite conditions."

The Northern flora has further undergone a specialization into three secondary floras, due to the combined influence of physical and genetic causes.

1. The Arctic-alpine flora (" consisting chiefly of plants of small stature, slow growth, and limited means of dispersion, compensated by long lives and great powers of endurance") is perhaps the most interesting of the three subdivisions, both because in its arctic aspect it reduces the divergence of the Old and New World divisions of the Northern flora to a minimum, and more especially on account of the great interest which attaches to the problem of its scattered alpine outliers. With regard to the first point, Hooker found that estimating the whole Arctic flora at 762 species, Arctic East America possessed 379, of which 269 were common to Scandinavia. Of the whole flora 616 species are found in Arctic Europe, and of these 586 are Scandinavian, and this leads Hooker to the striking observation that "the Scandinavian flora is present in every latitude of the globe, and is the only one that is so. "6 Christ objects to Hooker's giving the title of Scandinavian to the Arctic flora, but we must agree with Bentham that Scandinavia "would, according to older rules, have been regarded as the centre of creation for the arctic lands, and may now be termed the chief centre of preservation within the arctic circle, owing perhaps partly to its more broken conformation, and partly to that warmer climate which, while it now admits species which Christ objects to being included in the Arctic flora, was during the glacial period a means of preservation of some colder species which were everywhere else expelled or destroyed."

Just as at present the Arctic is more homogeneous than

the more southern divisions of the Northern flora, so we may infer that towards the close of the Tertiary epoch the continuous circumpolar land was covered with a vegetation also largely composed of identical plants but adapted to a warmer climate. As the climate became less warm there would commence a migration southwards, which would result in the modified descendants of these plants being now blended with the vegetation of Central Europe and the United States. As the glacial period gradually advanced, "the tropical plants and animals will have retreated from both sides towards the equator, followed in the rear by the temperate productions, and these by the arctic."1 When the climate of the earth again ameliorated, the migration took place in the reverse direction, and in this way mountain ranges became the havens of refuge of fragments of the original arctic floras which were exterminated on the lowlands. Even the equatorial region ceased to be a barrier during the glacial period, and to migration at that time must be attributed the survival of arctic forms in the south temperate zone. The southern migration of the Arctic flora does not appear to have taken place in one continuous wave. Thus, as Bentham points out,2 "many facts showed separate communications between the north and each of the three chains of the Pyrenees, the Alps, and the Himalayas, whilst these three gave little evidence of any lateral communication of their respective alpine vegetations."

The fact that the migration southwards and remigration northwards of the Arctic flora took place along parallels of longitude, accounts for some of its existing peculiarities. Hooker explains in this way the comparative poverty of

The Greenland flora.3

"If it be granted that the polar area was once occupied by the Scandinavian flora, and that the cold of the glacial epoch did drive this vegetation southwards, it is evident that the Greenland individuals, from being confined to a peninsula, would be exposed to very different conditions to those of the great continents. In Greenland many species would, as it were, be driven into the sea, that is exterminated; and the survivors would be confined to the southern portion of the peninsula, and not there being brought into competition with other types, there could be no struggle for life amongst their progeny, and consequently no selection of better adapted On the return of heat, these survivors would simply travel northwards unaccompanied by the plants of any other

varieties.

country.

"In Arctic America and Asia, on the other hand, where there Scandinavian plants to occupy, these would multiply enormously in individuals, branching off into varieties and sub-species, and occupy a larger area the further south they were driven; and none need be altogether lost in the southern migration over plains, though many would in the struggle that ensued when they reached the mountains of those continents and were brought into competition with the alpine plants, which the same cold had caused to descend to the plains. Hence, on the return of warmth, many more Scandinavian species would return to Arctic America and Asia than survived in Greenland; some would be changed in form, because only the favoured varieties could have survived the struggle; some of the alpine Siberian and Rocky Mountain species would accompany them to the arctic zone; while many arctic species would ascend those mountains, accompanying the alpine species in their

was a free southern extension and dilatation of land for the same

reascent."

A. de Candolle has very ingeniously applied the general principles laid down above to the detailed explanation of the distribution of the flora of the Alps themselves. The following is a brief summary of his conclusions :

The valleys and groups of mountains which have at present a maximum of rare species and the most varied flora belong to districts on which the glaciers disappeared earliest. On the other hand, where the duration of snows and glaciers has been most pro longed, the existing flora is poor. From a variety of causes which A. de Candolle enumerates, it seems probable that the southern and eastern glaciers of the Alps were of smaller extent than the northern, and would consequently be the soonest to retreat. We have consequently the curious fact that some of the most ancient southern slopes of the Alps. This is the case with species of fragments of the alpine flora are now only to be found on the Primula, Pedicularis, and Oxytropis, which exist neither in the interior of Switzerland, nor in the north of Europe. But it is easy to see that, like the other members of this flora, they were reappeared from underneath their snowy covering, while on the driven south during the glacial period, returning as the mountains northern side they were in great measure exterminated. Candolle points out as a fact in further confirmation that the Alpine species of Campanula, peculiar to Mont Cenis and the Simplon and find their nearest allies in Greece, Asia Minor, and the Himalaya. neighbouring valleys, are not related to the Arctic species, but

A. de

A further indication of the great antiquity of the Arcticalpine flora is afforded by the fact of its absence in the comparatively modern volcanic mountains of France. offer scarcely any of those alpine and sub-alpine plants "The Monts d'Or and Cantal, at an elevation of 6000 feet, which abound at the same or lower elevations in the Pyrenees on the one side, and in the Alps on the other, as well as in the British and Scandinavian mountains to the north."5 Hooker, however, points out that the absence of the alpine-arctic flora in Auvergne may be due to severe glaciation rather than to its absence (see Nature, Nov. 11, 1875, pp. 31, 32).

2. The Intermediate or Temperate flora is best described in the words of Bentham as

"A mongrel vegetation of mixed origin, including a large pro portion of species of the most extended geographical range, with a very few local ones, and those chiefly in the extreme west. The majority, whether trees, shrubs, or herbs, are plants of compara tively rapid growth, very prolific, endowed with great facilities for dispersion, and constitutions capable of adapting themselves to a great variety of physical and climatological conditions. They are great travellers, and soon take possession of any district left denuded by the abandonment of cultivation. To the great majority of them no primeval antiquity can be ascribed in Central or Western Europe; they appear to have come from the east, a considerable number perhaps from Western Asia, where their types appear to be more varied, but many also must have made half the tour of the globe. Large American genera have sent out offsets into Eastern Asia, which gradually diminishing in number of species, and sometimes slightly modifying their character, have spread over the whole of Asia, and invaded almost every part of Europe. These plants are, moreover, generally continuous, that is, interrupted only by intervals which under present condition they have means of crossing; and they are abundant in individuals, ascending in latitude and elevation, or descending to the south, until checked in their career by competing species, better enabled to endure the increasing rigour or the searching drought of the respective climates. Many of them will even assume slight modifications suited to their exceptional circumstances, and it is then as difficult to separate them from the genuine northern or southern floras as in many cases to give plausible grounds for establishing the precise origin of individual species."

דיי

The peat deposits of Denmark tell an unmistakable tale of the gradual advance of successive waves of vegetation from the south-east. The Scotch fir was once abundant within

p. 516.

These are given in greater fulness in Nature, April 27,,1876,
Bentham, Nat. Hist. Rev., 1864, p. 370.

Martins, however, considers that many of the plants of the

situations; thus the Oleander (Nerium Oleander) has been found in deposits from the Eocene upwards (Mém. de l'Acad. d. Sc. de Montpellier, ix. p. 95).

7 Nat. Hist. Rev., 1864, pp. 370, 971.

the Roman period in the Danish islands, but is now extinct; it was succeeded by the sessile-fruited oak, to be in turn supplanted by the pedunculated form of the same tree, associated with the alder, birch, and hazel. The oak is now almost supplanted by the beech. According to Areschoug, the original post-glacial flora of Scandinavia has retreated to the north, and is probably still retreating, while the flora of central and south Scandinavia consists of "an eastern and north-eastern vegetation, which spread into Europe after the glacial period and before the beech tree had invaded Sweden, with the admixture of more southern species, which, with the beech, have since penetrated into Sweden through Denmark."2 The beech and the chestnut occur in Japan, and, as far as Europe is concerned, there is good reason to regard their origin as Eastern.

As already pointed out, the American element in the European flora suffered severely during the glacial period, and has never since recovered itself. Japan, however, appeared to have been a great centre of preservation, and hence the numerous points of contact which its flora presents with that of the North American continent. In the New World itself, the continuity of the pre-glacial and post-glacial temperate floras has been better preserved. The following passage from an address of Asa Gray's may be quoted as giving its history in a concise form :

He "considered that the present vegetation or its proximate ancestry must have occupied the arctic and sub-arctic regions in Pliocene times, and that it had been gradually pushed southward as the temperature lowered and the glaciation advanced, even beyond its present habitation; that plants of the same stock and kindred probably ranging round the arctic zone as the present arctic species do, made their forced migration southwards upon widely different longitudes, and receded more or less as the climate grew warmer; that the general difference of climate which marks the eastern and western sides of the continents, -the one extreme, the other mean,—was doubtless even then established, so that the same species and the same sorts of species would be likely to secure and retain foothold in the similar climates of Japan and the Atlantic United States, but not in intermediate regions of different distribution of heat and moisture; so that different species of the same genus, as in Torreya, or different genera of the same group as redwood, Taxodium and Glyptostrobus, or different associations of foresttrees, might establish themselves each in the region best suited to the particular requirements, while they would fail to do so in any other."3

The west of Europe possesses the remains of a local and probably more ancient flora of very great interest, characterized by Gorse, and allied shrubby Leguminosa, Heaths, Lobelias, Sibthorpias, &c. These are closely checked in any tendency towards eastern dispersion by the severity of the winter climate away from the ameliorating influence of the sea. The probability of a southern extratropical connection of this peculiar element in the Northern flora will be adverted to hereafter.

The flora of the British Isles is in any respects interesting; it is in its main features an extension of the Germanic area of the temperate flora with the presence of the western element above alluded to distinctly marked on the south western coasts. Eriocaulon septangulare is an anomalous constituent, being limited to Ireland and a few islets on the western side of North Britain, and being otherwise an American and not a European species. Its presence can hardly be explained except by the agency of migratory birds.

3. The Mediterraneo-Caucasian flora, like the Arcticalpine, contrasts in the most marked way with the temperate.

"By far the richest and most diversified in species [it comprises six-sevenths of the European flora], it is also remarkable for the

1 Lyell. Antiquity of Man, p. 9.

2 Bentham. . c. p. 22

Darwiniana, pp. 224, 225.
Watson, Compendium, p. 31.

great variations centering round individual types, as well as for the very restricted areas occupied by a number of the most marked species; the limits are not to be accounted for by any physical peculiarities we are acquainted with, nor perhaps to be otherwise explained than by a supposition of very great antiquity."

Eastward of the Caucasus this remarkable flora dies away, reaching its eastern limit in Scinde, and the temperate flora of Asia is only separated from the tropical by the Himalayas. Southwards its progress is arrested by the arid zone formed by the African and Arabian deserts." As in the case of the Arctic flora, traces still exist of its former southern extension under the influence of a colder terrestrial climate. Adenocarpus, a characteristic Mediterranean genus, is represented by an identical species on Kilima Njaro, near the equator, and on the Cameroons mountains, 2000 miles distant on the opposite and western side of the African continent.6

II, THE SOUTHERN FLORA.-The Southern flora exhibits relations much more complex than those presented by the Northern. Instead of extending over large continental areas it is now dismembered into isolated groups scattered over the southern hemisphere, and in both the New and Old World sending northern extensions across the equator.

Five types may be briefly described, the definition of all but the first being taken from Bentham :-7

1. The Antarctic-alpine flora is the complement of the Arctic-alpine. It consists mainly of some widely distri buted northern genera such as Carex, Poa, Ranunculus, &c., with alpine types of strictly south temperate genera characteristic of the respective localities. Hooker describes it as possessing "decided Australian representatives in Centrolepideace and Stylidie, commencing in Fuegia, the Falklands, and Lord Auckland's and Campbell's groups, reappearing in the Alps of New Zealand, Tasmania, and Australia, and disappearing under the equator, on the Alps of Borneo."s

2. The Australian flora is "almost endemic, showing some slight connection with the New Zealand, and a few remains of former ramifications northward to some parts of the Indian Archipelago, a very few species, perhaps of modern introduction, extending to China and Japan." Bentham conclusively dismisses Unger's theory of the former extension of the Australian flora into Europe in Eocene times.

3. The Andine flora, characterized by a large number of distinct genera, Fuchsia, Gaultheria, Calceolaria, ranges more or less along the whole chain, "penetrating far northwards in Western America, throwing off a few brauches into Eastern Asia, and at its southern extremity crossing over to New Zealand, and in smaller numbers to Tasmania, and the mountains of Victoria." "10

4. The Mexico-Californian flora is "represented at great distances by closely allied species of small distinct genera -in Mexico and California, in the Argentine states, and in S. Africa or Australia."11

5. The S. African flora is "perhaps the richest known in proportion to its extent, and remarkably varied within its narrow limits." Its connection with other floras is very slight. That with Australia, alluded to at the commencement, does not extend beyond groups of the highest order (in the Proteaceae not merely the species but the genera

Bentham, Nat. Hist. Rev., 1864, p. 373. Hooker in Journ. Linn. Soc. Bot., xiv. p. 144. 'Presidential address, 1869, pp. 24, 25.

8 Introductory Essay to the Flora of Tasmania, p. 104. Presidential address, 1870, pp. 12-67.

10 On the extra-tropical southern connection between America and the Old World as illustrated by the Composita, see Bentham in Journ. Linn. Soc., xiii. p. 561.

11 See also Asa Gray, Darwiniana, pp. 218, 210.