My dear Lyell

I will have the following tremendous letter copied to make the reading easier, & as I want to keep a copy.—²

As you say you would like to hear my reasons for being most unwilling to believe in the continental extensions of late authors, I gladly write them; as, without I am convinced of my error, I shall have to give them condensed in my Essay, when I discuss single and multiple creation.³ I shall therefore be particularly glad to have your general opinion on them. I may quite likely have persuaded myself in my wrath that there is more in them than there is.— If there was much more reason to admit a continental extension in any one or two instances (as in Madeira) then in other cases, I should feel no difficulty whatever. But if on account of European plants and littoral sea-shells, it is thought necessary to join Madeira to mainland, Hooker is quite right to join New Holland to New Zealand, and Aukland Isl^d (and Raoul Is^d to N.E) and these to S. America and the Falklands, and these to Tristan d’Acunha, and these to Kerguelen Land;—thus making, either strictly at the same time, or at different periods, but all within the life of recent beings an almost circumpolar belt of land.⁴ So again Galapagos and Juan Fernandez must be joined to America; and if we trust to littoral sea-shells, the Galapagos must have been joined to Pacific isl^d. (2400 miles distant, as well as to America; and as Woodward seems to think all the islands in the Pacific into a magnificent continent:⁵ also the islands in the Southern Indian Ocean into another continent with Madagascar and Africa and perhaps India. In the N. Atlantic, Europe will stretch half way across the ocean to the Azores, and further North, right across. In short we must suppose probably, half the present ocean was land within the period of living organisms.— The globe within this period must have had a quite different aspect. Now the only way to test this, that I can see, is to consider whether the continents have undergone within this same period such wonderful permutations. In all N. and S. and central America, we have both recent and miocene (or eocene) shells, quite distinct on the opposite sides, and hence I cannot doubt that fundamentally America has held its place since at least the miocene period. In Africa almost all the living shells are distinct on the opposite sides of the intertropical regions, short as the distance is compared to the range of marine mollusca in uninterrupted seas; hence I infer that Africa has existed since our present species were created. Even the isthmus of Suez, and the Aralo-Caspian basin have had a great antiquity. So I imagine from the Tertiary deposits, has India. In Australia the great Fauna of extinct Marsupials, shows that before the present mammals appeared, Australia was a separate continent. I do not for one second doubt that very large portions of all these continents have undergone great changes of level within this period, but yet I conclude that fundamentally they stood as barriers in the sea, where they now stand; and therefore I should require the weightiest evidence to make me believe in such immense changes within the period of living organisms in our oceans; where, moreover from the great depths, the changes must have been vaster in a vertical sense.

Secondly. Submerge our present continents, leaving a few mountain peaks as islands, and what will the character of the islands be.— Consider that the Pyrenees, Sierra Nevada, Apennines, Alps, Carpathians are non-volcanic Etna and Caucausus, volcanic.— In Asia, Altai and Himmalaya, I believe non-volcanic.— In N. Africa the non-volcanic as I imagine alps of Abyssinia and of the Atlas: in S. Africa the Snow Mountains.— In Australia the non-volcanic Alps. In N. America, the White Mountains, Alleghanies, and Rocky Mountains, some of the latter alone I believe volcanic. In S. America to the East, the non-volcanic Silla of Caraccas and Itacolumi of Brazil, further South the S. Ventana; and in the Cordilleras, many volcanic but not all.— Now compare these peaks with the oceanic islands; as far as known all are volcanic, except S^t. Pauls (a strange bedevilled rock) and the Seychelles, if this latter can be called oceanic in the line of Madagascar; the Falklands only 500 miles off are only a shallow bank; New Caledonia, hardly oceanic, is another exception.— This argument has to me great weight. Compare on a Geographical⁶ map, islands which, we have several reasons to suppose were connected with mainland, as Sardinia and how different it appears. Believing, as I am inclined, that continents as continents, and oceans as oceans, are of immense antiquity, I should say that if the existing oceanic islands have any relation of any kind to continents, they are forming continents; and that by the time they could form a continent the volcanos would be denuded to their cores, leaving peaks of syenite diorite or porphyry. But have we nowhere any last wreck of a continent in the midst of the ocean? S^t. Pauls rock and such are old battered volcanic islands, as S^t. Helena may be; but I think we can see some reason why we should have less evidence of sinking than of rising continents (if my view in my Coral volume has any truth in it, viz: that volcanic outbursts accompany rising areas) for during subsidence there will be no compensating agent at work,—in rising areas there will be the additional element of out-poured volcanic matter.

Thirdly. Considering the depth of ocean, I was, before I got your letter, inclined vehemently to dispute the vast amount of subsidence, but I must strike my colours,— with respect to coral-reefs I carefully guarded against its being supposed that a continent was indicated by the groups of atolls.⁷ It is difficult to guess, as it seems to me, the amount of subsidence indicated by coral-reefs; but in such large areas, as the Low Arch: the Marshall Arch. and Laccadive group, it would, judging from the heights of existing oceanic archipelagoes, be odd if some peaks of from 8000 to 10,000 had not been buried.— Even after your letter a suspicion crossed me whether it would be fair to argue from subsidences in middle of greatest oceans to continents; but refreshing my memory by talking with Ramsay in regard to probable thickness in one vertical line of the Silurian and carboniferous formations,⁸ it seems there must have been at least 10,000 feet of subsidence during these formations in Europe and N. America, and therefore during the continuance of nearly the same set of organic beings. But even 12,000 feet would not be enough for the Azores or for Hooker’s continent. I believe Hooker does not infer a continuous continent, but approximate groups of islands, with, if we may judge from existing continents, not profoundly deep sea between them: but the argument from the volcanic nature of nearly every existing oceanic island tell against such supposed groups of islands,—for I presume he does not suppose a mere chain of volcanic islands belting the Southern Hemisphere.

Fourthly.— The supposed continental extensions do not seem to me perfectly to account for all the phenomena of distribution on islands:—as the absence of mammals and Batrachians; the absence of certain great groups of insects on Madeira, and of Acaciæ and Banksias &c in New Zealand:—the paucity of plants in some cases &c.— Not that those who believe in various accidental means of dispersal can explain most of these cases; but they may at least say that these facts seem hardly compatable with former continuous land.

Finally, for these several reasons, and especially considering it certain (in which you will agree) that we are extremely ignorant of means of dispersal, I cannot avoid thinking that Forbe’s Atlantis was an ill-service to Science, as checking a close study of means of dissemination. I shall be really grateful to hear as briefly as you like, whether these arguments have any weight with you, putting yourself in the position of an honest judge. I told Hooker I was going to write to you on this subject; and I should like him to read this; but whether he or you will think it worth time and postage remains to be proved.—

Yours most truly. | Charles Darwin