If we now look at the distribution of sea-shells, we find the same changes in distribution. The Red Sea and the Mediterranean were more nearly related in these shells than they now are. In different parts of Europe, on the other hand, during the Miocene period, the sea-shells seem to have been more different than at present. In{395} the Tertiary period, according to Lyell, the shells of N. America and Europe were less related than at present, and during the Cretaceous still less like; whereas, during this same Cretaceous period, the shells of India and Europe were more like than at present. But going further back to the Carbonaceous period, in N. America and Europe, the productions were much more like than they now are{396}. These facts harmonise with the conclusions drawn from the present distribution of organic beings, for we have seen, that from species being created in different points or areas, the formation of a barrier would cause or make two distinct geographical areas; and the destruction of a barrier would permit their diffusion{397}. And as long-continued geological changes must both destroy and make barriers, we might expect, the further we looked backwards, the more changed should we find the present distribution. This conclusion is worthy of attention; because, finding in widely different parts of the same main division of the world, and in volcanic islands near them, groups of distinct, but related, species;--and finding that a singularly analogous relation holds good with respect to the beings of past times, when none of the present species were living, a person might be tempted to believe in some mystical relation between certain areas of the world, and the production of certain organic forms; but we now see that such an assumption would have to be complicated by the admission that such a relation, though holding good for long revolutions of years, is not truly persistent. {395} <Note by the author.> All this requires much verification. {396} This point seems to be less insisted on in the _Origin_. {397} _Origin_, Ed. i. p. 356, vi. p. 504.