Long after this the Isle of Wight was altogether separated by the Solent from the mainland, but still ages before the historic period. The various traditions, as to the former depth of the channel, how Sir Bevis, of Southampton, waded across it, how, too, the carts brought the Binstead stone for building Beaulieu Abbey over the dry bed at low water, have been previously given. The passage, too, in Diodorus Siculus has been already examined,[265] and there can be no doubt, notwithstanding his also making it, like the traditions, a peninsula at low water, that his Ictis is the Isle of Wight and not St. Michael’s Mount. The mere local evidence of the mass of tin, the British road—more like a deep trench than a road—still plainly traceable across the Forest, the names along it corresponding with that of its continuation in the Island, would alone, most assuredly, show that this was the place whence the first traders, and, in after-times, the Romans, exported their tin. We must, however, remember that the channel of the Solent was caused by depression rather than by excavation; and that at this moment an alteration in the levels, as noticed by Mr. Austen,[266] is going on eastward of Hurst Castle. The drift, which spreads over the whole of the New Forest, is not very interesting. No elephants’ tusks, or elks’ horns, so far as I know, have ever been discovered. A few species of _Terebratula_ and _Pecten_, some flint knives, and the _os inominatum_, of probably _Bos longifrons_, mentioned farther on, are the only things at present found. Still, in one way, it is most interesting, as completely disproving the Chroniclers’ accounts that, before its afforestation by the Conqueror, the district of the Forest was so fertile. The fact is a sheer impossibility. No wheat could ever be grown on this great bed of chalk-gravel, which is varied only by patches of sand. But nowhere, perhaps, in the world can we see the stratification of the upper portion of the Middle-Eocene better than at Hordle and Barton, as the sea serves to keep the different strata exposed. The beds dip easterly with a fall of about one in a hundred, though, at the extreme west, at High Cliff, it is much less, and here and there in some few places they lie almost horizontally.[267] At Hordle they seem to have been deposited in a river of a very uniform depth. There is but one single fault in the whole series, just under Mead End, where all the beds have alike suffered. Here and there, however, they are deposited with an undulating line; and here and there, too, a rippled surface occurs, caused by the action of small waves. The river appears to have varied very much in the amount and force of its stream, as some of the beds, where the shells are less frequent, have been deposited very rapidly, whilst others, where the organic remains are more abundant, have been laid on very slowly and in very still water.[268]