This is considerably larger than the other two great outlets of the subterranean waters of Mendip, those of Wookey Hole and Wells, each of which, however, pours forth an enormous volume. That it is the Cheddar stream which is responsible for the existence of the gorge itself no one can doubt, and it is a most interesting subject for discussion as to how this has been brought about. It is not difficult to determine what points must mark the boundaries of the catchment area, the waters of which drain to Cheddar. The road from Castle Comfort to Charterhouse on the north-east, the outcrop of Shales south of Blackdown on the north, and a line drawn from Rowberrow Farm north of Priddy to the gorge itself on the south, enclose the whole area from which the supply is obtained. This is somewhere about 12 square miles in extent. To this must be added, possibly, some water from slightly more to the eastward. It is now the commonly accepted theory that the whole of this water, or at any rate the bulk of it, found inlet into a series of caverns along the line now occupied by the gorge, and that then the processes which are so well known to be going on gradually enlarged these to the point of collapse, the falling débris being removed by the still flowing stream. It is only right to add that M. Martel, arguing from his long experience, which probably exceeds that of any man who has ever studied the subject, sees in the gorges of Cheddar, Burrington, and presumably Ebbor, the superficial channels worn by the escaping streams from the ancient Mendip plateau. He says, "The numerous dried valleys (Burrington Combe, Cheddar Cliffs, etc.), which cut through the circumference of the Mendips, witness, as everywhere, to the ancient superficial flowing off of the rivers, and to their capture by the natural wells, successively opened and enlarged in the cracks of the Limestone rock." That even small streams acting through a sufficient period of time are capable of doing enormous erosive work it would be idle to deny, but the difficulties in the way of accepting this theory as alone sufficient are too great to admit of its acceptance. It demands that the water of a very large area could find access to the eastern end of the ravine, which itself demands that the general configuration of the Mendips must have been very different from that presented now. This, from the existence of the Secondary beds in their present position, say near Harptree, was not the case; and therefore, for the theory to hold good, we must suppose that the superficial gorge was pre-Triassic. As it was not filled in, either in Triassic time or subsequently, it could not have been superficial. Of course it may be contended that the reversal of this line of argument demonstrates that the gorge is post-Liassic and may then have been a superficial channel, but I hold this to be disproved in my chapter on the antiquity of the Mendip Caves. I am, accordingly, forced to the conclusion that the