Inspection of a map of the world at once shows that the five divisions, separated according to the greatest amount of difference in the mammifers inhabiting them, are likewise those most widely separated from each other by barriers{345} which mammifers cannot pa.s.s: thus Australia is separated from New Guinea and some small adjoining islets only by a narrow and shallow strait; whereas New Guinea and its adjoining islets are cut off from the other East Indian islands by deep water. These latter islands, I may remark, which fall into the great Asiatic group, are separated from each other and the continent only by shallow water; and where this is the case we may suppose, from geological oscillations of level, that generally there has been recent union. South America, including the southern part of Mexico, is cut off from North America by the West Indies, and the great table-land of Mexico, except by a mere fringe of tropical forests along the coast: it is owing, perhaps, to this fringe that N. America possesses some S. American forms. Madagascar is entirely isolated. Africa is also to a great extent isolated, although it approaches, by many promontories and by lines of shallower sea, to Europe and Asia: southern Africa, which is the most distinct in its mammiferous inhabitants, is separated from the northern portion by the Great Sahara Desert and the table-land of Abyssinia. That the distribution of organisms is related to barriers, stopping their progress, we clearly see by comparing the distribution of marine and terrestrial productions. The marine animals being different on the two sides of land tenanted by the same terrestrial animals, thus the sh.e.l.ls are wholly different on the opposite sides of the temperate parts of South America{346}, as they are (?) in the Red Sea and the Mediterranean. We can at once perceive that the destruction of a barrier would permit two geographical groups of organisms to fuse and blend into one. But the original cause of groups being different on opposite sides of a barrier can only be understood on the hypothesis of each organism having been created or produced on one spot or area, and afterwards migrating as widely as its means of transport and subsistence permitted it.

{345} On the general importance of barriers, see _Origin_, Ed. i.

p. 347, vi. p. 494.

{346} _Origin_, Ed. i. p. 348, vi. p. 495.

_Relation of range in genera and species._

It is generally{347} found, that where a genus or group ranges over nearly the entire world, many of the species composing the group have wide ranges: on the other hand, where a group is restricted to any one country, the species composing it generally have restricted ranges in that country{348}. Thus among mammifers the feline and canine genera are widely distributed, and many of the individual species have enormous ranges [the genus Mus I believe, however, is a strong exception to the rule]. Mr Gould informs me that the rule holds with birds, as in the owl genus, which is mundane, and many of the species range widely. The rule holds also with land and fresh-water mollusca, with b.u.t.terflies and very generally with plants. As instances of the converse rule, I may give that division of the monkeys which is confined to S. America, and amongst plants, the Cacti, confined to the same continent, the species of both of which have generally narrow ranges. On the ordinary theory of the separate creation of each species, the cause of these relations is not obvious; we can see no reason, because many allied species have been created in the several main divisions of the world, that several of these species should have wide ranges; and on the other hand, that species of the same group should have narrow ranges if all have been created in one main division of the world. As the result of such and probably many other unknown relations, it is found that, even in the same great cla.s.ses of beings, the different divisions of the world are characterised by either merely different species, or genera, or even families: thus in cats, mice, foxes, S. America differs from Asia and Africa only in species; in her pigs, camels and monkeys the difference is generic or greater. Again, whilst southern Africa and Australia differ more widely in their mammalia than do Africa and S. America, they are more closely (though indeed very distantly) allied in their plants.

{347} The same laws seem to govern distribution of species and genera, and individuals in time and s.p.a.ce. {348} _Origin_, Ed. i. p. 404, vi. p. 559._Distribution of the inhabitants in the same continent._ If we now look at the distribution of the organisms in any one of the above main divisions of the world, we shall find it split up into many regions, with all or nearly all their species distinct, but yet partaking of one common character. This similarity of type in the subdivisions of a great region is equally well-known with the dissimilarity of the inhabitants of the several great regions; but it has been less often insisted on, though more worthy of remark. Thus for instance, if in Africa or S. America, we go from south to north{349}, or from lowland to upland, or from a humid to a dryer part, we find wholly different species of those genera or groups which characterise the continent over which we are pa.s.sing. In these subdivisions we may clearly observe, as in the main divisions of the world, that sub-barriers divide different groups of species, although the opposite sides of such sub-barriers may possess nearly the same climate, and may be in other respects nearly similar: thus it is on the opposite sides of the Cordillera of Chile, and in a lesser degree on the opposite sides of the Rocky mountains. Deserts, arms of the sea, and even rivers form the barriers; mere preoccupied s.p.a.ce seems sufficient in several cases: thus Eastern and Western Australia, in the same lat.i.tude, with very similar climate and soils, have scarcely a plant, and few animals or birds, in common, although all belong to the peculiar genera characterising Australia. It is in short impossible to explain the differences in the inhabitants, either of the main divisions of the world, or of these sub-divisions, by the differences in their physical conditions, and by the adaptation of their inhabitants. Some other cause must intervene.{349} _Origin_, Ed. i. p. 349, vi. p. 496.We can see that the destruction of sub-barriers would cause (as before remarked in the case of the main divisions) two sub-divisions to blend into one; and we can only suppose that the original difference in the species, on the opposite sides of sub-barriers, is due to the creation or production of species in distinct areas, from which they have wandered till arrested by such sub-barriers. Although thus far is pretty clear, it may be asked, why, when species in the same main division of the world were produced on opposite sides of a sub-barrier, both when exposed to similar conditions and when exposed to widely different influences (as on alpine and lowland tracts, as on arid and humid soils, as in cold and hot climates), have they invariably been formed on a similar type, and that type confined to this one division of the world?Why when an ostrich{350} was produced in the southern parts of America, was it formed on the American type, instead of on the African or on Australian types? Why when hare-like and rabbit-like animals were formed to live on the Savannahs of La Plata, were they produced on the peculiar Rodent type of S. America, instead of on the true{351} hare-type of North America, Asia and Africa? Why when borrowing Rodents, and camel-like animals were formed to tenant the Cordillera, were they formed on the same type{352} with their representatives on the plains?Why were the mice, and many birds of different species on the opposite sides of the Cordillera, but exposed to a very similar climate and soil, created on the same peculiar S. American type? Why were the plants in Eastern and Western Australia, though wholly different as species, formed on the same peculiar Australian types? The generality of the rule, in so many places and under such different circ.u.mstances, makes it highly remarkable and seems to demand some explanation.{350} The case of the ostrich (_Rhea_) occurs in the _Origin_, Ed.i. p. 349, vi. p. 496.{351} There is a hare in S. America,--so bad example.{352} See _Origin_, Ed. i. p. 349, vi. p. 497._Insular Faunas._ If we now look to the character of the inhabitants of small islands{353}, we shall find that those situated close to other land have a similar fauna with that land{354}, whilst those at a considerable distance from other land often possess an almost entirely peculiar fauna. The Galapagos Archipelago{355} is a remarkable instance of this latter fact; here almost every bird, its one mammifer, its reptiles, land and sea sh.e.l.ls, and even fish, are almost all peculiar and distinct species, not found in any other quarter of the world: so are the majority of its plants. But although situated at the distance of between 500 and 600 miles from the S. American coast, it is impossible to even glance at a large part of its fauna, especially at the birds, without at once seeing that they belong to the American type{356}. Hence, in fact, groups of islands thus circ.u.mstanced form merely small but well-defined sub-divisions of the larger geographical divisions. But the fact is in such cases far more striking: for taking the Galapagos Archipelago as an instance; in the first place we must feel convinced, seeing that every island is wholly volcanic and bristles with craters, that in a geological sense the whole is of recent origin comparatively with a continent; and as the species are nearly all peculiar, we must conclude that they have in the same sense recently been produced on this very spot; and although in the nature of the soil, and in a lesser degree in the climate, there is a wide difference with the nearer part of the S.American coast, we see that the inhabitants have been formed on the same closely allied type. On the other hand, these islands, as far as their physical conditions are concerned, resemble closely the Cape de Verde volcanic group, and yet how wholly unlike are the productions of these two archipelagoes. The Cape de Verde{357} group, to which may be added the Canary Islands, are allied in their inhabitants (of which many are peculiar species) to the coast of Africa and southern Europe, in precisely the same manner as the Galapagos Archipelago is allied to America. We here clearly see that mere geographical proximity affects, more than any relation of adaptation, the character of species. How many islands in the Pacific exist far more like in their physical conditions to Juan Fernandez than this island is to the coast of Chile, distant 300 miles; why then, except from mere proximity, should this island alone be tenanted by two very peculiar species of humming-birds--that form of birds which is so exclusively American? Innumerable other similar cases might be adduced.{353} For the general problem of Oceanic Islands, see _Origin_, Ed.i. p. 388, vi. p. 541.{354} This is an ill.u.s.tration of the general theory of barriers (_Origin_, Ed. i. p. 347, vi. p. 494). At i. p. 391, vi. p. 544 the question is discussed from the point of view of means of transport.Between the lines, above the words "with that land," the author wrote "Cause, formerly joined, no one doubts after Lyell."{355} _Origin_, Ed. i. p. 390, vi. p. 543.{356} See _Origin_, Ed. i. p. 397, vi. p. 552.{357} The Cape de Verde and Galapagos Archipelagoes are compared in the _Origin_, Ed. i. p. 398, vi. p. 553. See also _Journal of Researches_, 1860, p. 393.The Galapagos Archipelago offers another, even more remarkable, example of the cla.s.s of facts we are here considering. Most of its genera are, as we have said, American, many of them are mundane, or found everywhere, and some are quite or nearly confined to this archipelago.The islands are of absolutely similar composition, and exposed to the same climate; most of them are in sight of each other; and yet several of the islands are inhabited, each by peculiar species (or in some cases perhaps only varieties) of some of the genera characterising the archipelago. So that the small group of the Galapagos Islands typifies, and follows exactly the same laws in the distribution of its inhabitants, as a great continent. How wonderful it is that two or three closely similar but distinct species of a mocking-thrush{358} should have been produced on three neighbouring and absolutely similar islands; and that these three species of mocking-thrush should be closely related to the other species inhabiting wholly different climates and different districts of America, and only in America. No similar case so striking as this of the Galapagos Archipelago has. .h.i.therto been observed; and this difference of the productions in the different islands may perhaps be partly explained by the depth of the sea between them (showing that they could not have been united within recent geological periods), and by the currents of the sea sweeping _straight_ between them,--and by storms of wind being rare, through which means seeds and birds could be blown, or drifted, from one island to another. There are however some similar facts: it is said that the different, though neighbouring islands of the East Indian Archipelago are inhabited by some different species of the same genera; and at the Sandwich group some of the islands have each their peculiar species of the same genera of plants.{358} In the _Origin_, Ed. i. p. 390, a strong point is made of birds which immigrated "with facility and in a body" not having been modified. Thus the author accounts for the small percentage of peculiar "marine birds."Islands standing quite isolated within the intra-tropical oceans have generally very peculiar floras, related, though feebly (as in the case of St Helena{359} where almost every species is distinct), with the nearest continent: Tristan d'Acunha is feebly related, I believe, in its plants, both to Africa and S. America, not by having species in common, but by the genera to which they belong{360}. The floras of the numerous scattered islands of the Pacific are related to each other and to all the surrounding continents; but it has been said, that they have more of an Indo-Asiatic than American character{361}. This is somewhat remarkable, as America is nearer to all the Eastern islands, and lies in the direction of the trade-wind and prevailing currents; on the other hand, all the heaviest gales come from the Asiatic side. But even with the aid of these gales, it is not obvious on the ordinary theory of creation how the possibility of migration (without we suppose, with extreme improbability, that each species with an Indo-Asiatic character has actually travelled from the Asiatic sh.o.r.es, where such species do not now exist) explains this Asiatic character in the plants of the Pacific. This is no more obvious than that (as before remarked) there should exist a relation between the creation of closely allied species in several regions of the world, and the fact of many such species having wide ranges; and on the other hand, of allied species confined to one region of the world having in that region narrow ranges.{359} "The affinities of the St Helena flora are strongly South African." Hooker's _Lecture on Insular Floras_ in the _Gardeners'Chronicle_, Jan. 1867.{360} It is impossible to make out the precise form which the author intended to give to this sentence, but the meaning is clear.{361} This is no doubt true, the flora of the Sandwich group however has marked American affinities._Alpine Floras._ We will now turn to the floras of mountain-summits which are well known to differ from the floras of the neighbouring lowlands. In certain characters, such as dwarfness of stature, hairiness, &c., the species from the most distant mountains frequently resemble each other,--a kind of a.n.a.logy like that for instance of the succulency of most desert plants. Besides this a.n.a.logy, Alpine plants present some eminently curious facts in their distribution. In some cases the summits of mountains, although immensely distant from each other, are clothed by the same identical species{362} which are likewise the same with those growing on the likewise very distant Arctic sh.o.r.es. In other cases, although few or none of the species may be actually identical, they are closely related; whilst the plants of the lowland districts surrounding the two mountains in question will be wholly dissimilar. As mountain-summits, as far as their plants are concerned, are islands rising out of an ocean of land in which the Alpine species cannot live, nor across which is there any known means of transport, this fact appears directly opposed to the conclusion which we have come to from considering the general distribution of organisms both on continents and on islands--namely, that the degree of relationship between the inhabitants of two points depends on the completeness and nature of the barriers between those points{363}. I believe, however, this anomalous case admits, as we shall presently see, of some explanation. We might have expected that the flora of a mountain summit would have presented the same relation to the flora of the surrounding lowland country, which any isolated part of a continent does to the whole, or an island does to the mainland, from which it is separated by a rather wide s.p.a.ce of sea.This in fact is the case with the plants clothing the summits of _some_ mountains, which mountains it may be observed are particularly isolated; for instance, all the species are peculiar, but they belong to the forms characteristic of the surrounding continent, on the mountains of Caraccas, of Van Dieman's Land and of the Cape of Good Hope{364}. On some other mountains, for instance Tierra del Fuego and in Brazil, some of the plants though distinct species are S. American forms; whilst others are allied to or are identical with the Alpine species of Europe.In islands of which the lowland flora is distinct but allied to that of the nearest continent, the Alpine plants are sometimes (or perhaps mostly) eminently peculiar and distinct{365}; this is the case on Teneriffe, and in a lesser degree even on some of the Mediterranean islands.{362} See _Origin_, Ed. i. p. 365, vi. p. 515. The present discussion was written before the publication of Forbes' celebrated paper on the same subject; see _Life and Letters_, vol. I. p. 88.{363} The apparent breakdown of the doctrine of barriers is slightly touched on in the _Origin_, Ed. i. p. 365, vi. p. 515.{364} In the _Origin_, Ed. i. p. 375, vi. p. 526, the author points out that on the mountains at the Cape of Good Hope "some few representative European forms are found, which have not been discovered in the inter-tropical parts of Africa."{365} See Hooker's _Lecture on Insular Floras_ in the _Gardeners'Chronicle_, Jan. 1867.If all Alpine floras had been characterised like that of the mountain of Caraccas, or of Van Dieman's Land, &c., whatever explanation is possible of the general laws of geographical distribution would have applied to them. But the apparently anomalous case just given, namely of the mountains of Europe, of some mountains in the United States (Dr Boott) and of the summits of the Himalaya (Royle), having many identical species in common conjointly with the Arctic regions, and many species, though not identical, closely allied, require a separate explanation.The fact likewise of several of the species on the mountains of Tierra del Fuego (and in a lesser degree on the mountains of Brazil) not belonging to American forms, but to those of Europe, though so immensely remote, requires also a separate explanation._Cause of the similarity in the floras of some distant mountains._ Now we may with confidence affirm, from the number of the then floating icebergs and low descent of the glaciers, that within a period so near that species of sh.e.l.ls have remained the same, the whole of Central Europe and of North America (and perhaps of Eastern Asia) possessed a very cold climate; and therefore it is probable that the floras of these districts were the same as the present Arctic one,--as is known to have been to some degree the case with then existing sea-sh.e.l.ls, and those now living on the Arctic sh.o.r.es. At this period the mountains must have been covered with ice of which we have evidence in the surfaces polished and scored by glaciers. What then would be the natural and almost inevitable effects of the gradual change into the present more temperate climate{366}? The ice and snow would disappear from the mountains, and as new plants from the more temperate regions of the south migrated northward, replacing the Arctic plants, these latter would crawl{367} up the now uncovered mountains, and likewise be driven northward to the present Arctic sh.o.r.es. If the Arctic flora of that period was a nearly uniform one, as the present one is, then we should have the same plants on these mountain-summits and on the present Arctic sh.o.r.es. On this view the Arctic flora of that period must have been a widely extended one, more so than even the present one; but considering how similar the physical conditions must always be of land bordering on perpetual frost, this does not appear a great difficulty; and may we not venture to suppose that the almost infinitely numerous icebergs, charged with great ma.s.ses of rocks, soil and _brushwood_{368} and often driven high up on distant beaches, might have been the means of widely distributing the seeds of the same species?{366} In the margin the author has written "(Forbes)." This may have been inserted at a date later than 1844, or it may refer to a work by Forbes earlier than his Alpine paper.{367} See _Origin_, Ed. i. p. 367, vi. p. 517.{368} Perhaps vitality checked by cold and so prevented germinating. I will only hazard one other observation, namely that during the change from an extremely cold climate to a more temperate one the conditions, both on lowland and mountain, would be singularly favourable for the diffusion of any existing plants, which could live on land, just freed from the rigour of eternal winter; for it would possess no inhabitants; and we cannot doubt that _preoccupation_{369} is the chief bar to the diffusion of plants. For amongst many other facts, how otherwise can we explain the circ.u.mstance that the plants on the opposite, though similarly const.i.tuted sides of a wide river in Eastern Europe (as I was informed by Humboldt) should be widely different; across which river birds, swimming quadrupeds and the wind must often transport seeds; we can only suppose that plants already occupying the soil and freely seeding check the germination of occasionally transported seeds.{369} A note by the author gives "many authors" apparently as authority for this statement.