Commencing with the most easterly of the rivers which enter the Sound, we find that the Laira Railway Viaduct, across the Cattewater, proved a breadth of 212 feet at the centre of the channel, with the rock-floor practically level at 87 feet below low water; no V-shaped valley or gorge was met with. At Saltash the foundations of the bridge show the depth to the rock-bottom to be 75 feet; but the viaduct across the Hamoaze is about three miles higher up the river than the Laira Viaduct. The Tavy Viaduct, nearly two miles further from the sea, shows a width of 240 feet of practically level rocky floor at 67 feet below sea-level. Thus all this evidence is consistent with the existence of a series of wide open flat-bottomed valleys, now partly submerged, with a fall of about five feet in the mile. This is about the fall necessary for even a rapid river flowing through a flat so full of boulders and coarse gravel as this must have been. It must not be forgotten also that this five feet in the mile is the general fall of the valley-bottom, not of the water, and that a river winding from side to side would have about half or one-third of this fall. The slope was probably just sufficient to keep the channel clear and let the water escape. We may take it, therefore, that the ancient valleys opening into Plymouth Harbour cut to about 100 feet below mean tide, as do the Thames and Humber, and that this was the measure of the greatest elevation of the land in Pleistocene times, for these valleys opened suddenly into a sea of considerably greater depth. A word of explanation is still required as to the meaning of the extremely flat rock-bottom, for one might have expected more of a =U=-shaped or =V=-shaped valley, unless the period of stationary sea-level were very long. Owing to the great rush of water from Dartmoor during floods, and the enormous amount of coarse gravel swept down, the erosive power of these streams is very great. This was greatly exaggerated during the Glacial Epoch, to which the formation of the tin-ground and of the flat bottom belong. The melting of the snow in spring must have caused far more severe floods than we now see, and these floods must have brought down large quantities of river-ice heavily charged with boulders of hard and angular metamorphic rocks, such as would erode and trench in a way that does not now happen. Thus as the river changed its course or swung from side to side according to the varying amount of water, the ice-laden water must have had an erosive power more like that of a Canadian river in spring than like anything we now see in Britain. The wide and deep flat-bottomed trench need not have taken any enormous length of time to form, for river-ice and anchor-ice were constantly at work removing the loose material and laying bare the rock-face so that it could be again attacked.