The first portion of the subject—that which relates to the various modes by which organisms can pass over wide areas of sea and land—has been fully treated by Sir Charles Lyell, by Mr. Darwin, and many other writers, and it will only be necessary here to give a very brief notice of the best known facts on the subject, which will be further referred to when we come to discuss the particular cases that arise in regard to the faunas and floras of remote islands. But the other side of the question of dispersal—that which depends on geological and climatal changes—is in a far less satisfactory condition, for, though much has been written upon it, the most contradictory opinions still prevail, and at almost every step we find ourselves on the battle-field of opposing schools in geological or physical science. As, however, these questions lie at the very root of any general solution of the problems of distribution, I have given much time to a careful examination of the various theories that have been advanced, and the discussions to which they have given rise; and have arrived at some definite conclusions which I venture to hope may serve as the foundation for a better comprehension of these intricate problems. The four chapters which follow this are devoted to a full examination of these profoundly interesting and important questions, after which we shall enter upon our special inquiry—the nature and origin of insular faunas and floras.
The Ocean as a Barrier to the Dispersal of Mammals.—A wide extent of ocean forms an almost absolute barrier to the dispersal of all land animals, and of most of those which are aerial, since even birds cannot fly for thousands of miles without rest and without food, unless they are aquatic birds which can find both rest and food on the surface of the ocean. We may be sure, therefore, that without artificial help neither mammalia nor land birds can pass over very wide oceans. The exact width they can pass over is not determined, but we have a few facts to guide us. Contrary to the common notion, pigs can swim very well, and have been known to swim over five or six miles of sea, and the wide distribution of pigs in the islands of the Eastern Hemisphere may be due to this power. It is almost certain, however, that they would never voluntarily swim away from their native land, and if carried out to sea by a flood they would certainly endeavour to return to the shore. We cannot therefore believe that they would ever swim over fifty or a hundred miles of sea, and the same may be said of all the larger mammalia. Deer also swim well, but there is no reason to believe that they would venture out of sight of land. With the smaller, and especially with the arboreal mammalia, there is a much more effectual way of passing over the sea, by means of floating trees, or those floating islands which are often formed at the mouths of great rivers. Sir Charles Lyell describes such floating islands which were encountered among the Moluccas, on which trees and shrubs were growing on a stratum of soil which even formed a white beach round the margin of each raft. Among the Philippine Islands similar rafts with trees growing on them have been seen after hurricanes; and it is easy to understand how, if the sea were tolerably calm, such a raft might be carried along by a current, aided by the wind acting on the trees, till after a passage of several weeks it might arrive safely on the shores of some land hundreds of miles away from its starting-point. Such small animals as squirrels and field-mice might have been carried away on the trees which formed part of such a raft, and might thus colonise a new island; though, as it would require a pair of the same species to be thus conveyed at the same time, such accidents would no doubt be rare. Insects, however, and land-shells would almost certainly be abundant on such a raft or island, and in this way we may account for the wide dispersal of many species of both these groups.
Notwithstanding the occasional action of such causes, we cannot suppose that they have been effective in the dispersal of mammalia as a whole; and whenever we find that a considerable number of the mammals of two countries exhibit distinct marks of relationship, we may be sure that an actual land connection, or at all events an approach to within a very few miles of each other, has at one time existed. But a considerable number of identical mammalian families and even genera are actually found in all the great continents, and the present distribution of land upon the globe renders it easy to see how they have been able to disperse themselves so widely. All the great land masses radiate from the arctic regions as a common centre, the only break being at Behrings Strait, which is so shallow that a rise of less than a thousand feet would form a broad isthmus connecting Asia and America as far south as the parallel of 60° N. Continuity of land therefore may be said to exist already for all parts of the world (except Australia and a number of large islands, which will be considered separately), and we have thus no difficulty in the way of that former wide diffusion of many groups, which we maintain to be the only explanation of most anomalies of distribution other than such as may be connected with unsuitability of climate.
The Dispersal of Birds.—Wherever mammals can migrate other vertebrates can generally follow with even greater facility. Birds, having the power of flight, can pass over wide arms of the sea, or even over extensive oceans, when these are, as in the Pacific, studded with islands to serve as resting places. Even the smaller land-birds are often carried by violent gales of wind from Europe to the Azores, a distance of nearly a thousand miles, so that it becomes comparatively easy to explain the exceptional distribution of certain species of birds. Yet on the whole it is remarkable how closely the majority of birds follow the same laws of distribution as mammals, showing that they generally require either continuous land or an island-strewn sea as a means of dispersal to new homes.
The Dispersal of Reptiles.—Reptiles appear at first sight to be as much dependent on land for their dispersal as mammalia, but they possess two peculiarities which favour their occasional transmission across the sea—the one being their greater tenacity of life, the other their oviparous mode of reproduction. A large boa-constrictor was once floated to the island of St. Vincent, twisted round the trunk of a cedar tree, and was so little injured by its voyage that it captured some sheep before it was killed. The island is nearly two hundred miles from Trinidad and the coast of South America, whence the reptile almost certainly came.[14] Snakes are, however, comparatively scarce on islands far from continents, but lizards are often abundant, and though these might also travel on floating trees, it seems more probable that there is some as yet unknown mode by which their eggs are safely, though perhaps very rarely, conveyed from island to island. Examples of their peculiar distribution will be given when we treat of the fauna of some islands in which they abound.
The Dispersal of Amphibia and Fresh-water Fishes.—The two lower groups of vertebrates, Amphibia and fresh-water fishes, possess special facilities for dispersal, in the fact of their eggs being deposited in water, and in their aquatic or semi-aquatic habits. They have another advantage over reptiles in being capable of flourishing in arctic regions, and in the power possessed by their eggs of being frozen without injury. They have thus, no doubt, been assisted in their dispersal by floating ice, and by that approximation of all the continents in high northern latitudes which has been the chief agent in producing the general uniformity in the animal productions of the globe. Some genera of Batrachia have almost a world-wide distribution; while the tailed Batrachia, such as the newts and salamanders, are almost entirely confined to the northern hemisphere, some of the genera spreading over the whole of the north temperate zone. Fresh-water fishes have often a very wide range, the same species being sometimes found in all the rivers of a continent. This is no doubt chiefly due to the want of permanence in river basins, especially in their lower portions, where streams belonging to distinct systems often approach each other and may be made to change their course from one to the other basin by very slight elevations or depressions of the land. Hurricanes and water-spouts also often carry considerable quantities of water from ponds and rivers, and thus disperse eggs and even small fishes. As a rule, however, the same species are not often found in countries separated by a considerable extent of sea, and in the tropics rarely the same genera. The exceptions are in the colder regions of the earth, where the transporting power of ice may have come into play. High ranges of mountains, if continuous for