Possibly a deep channel may exist towards the Atlantic; but we know that none extends as far up as Bristol. Near Bristol the Severn Tunnel was carried through Carboniferous and Triassic rock, and showed that no buried channel is found much below the present one, which here happens to be scoured by the tides to an exceptional depth. The bottom of the old channel cannot be more than 40 feet below the bottom of the present channel known as The Shoots. It may be that the Severn was once prolonged seaward as a swift river falling in a series of rapids over hard ledges of Palaeozoic rocks; but of this there is no evidence. It also does not seem probable, for all the geological indications go to suggest that west of Bristol the Channel coincides in the main with a wide area once occupied by comparatively soft Secondary or even Tertiary rocks. However this may be, we can only trace an ancient post-glacial channel cutting to about the same depth as the channels of the other rivers, and the lowest submerged land-surface of Barry Docks corresponds quite well with an alluvial flat formed when the river ran at that level. Here again we seem to find the river cutting to an ancient base-level which was about 60 feet below the present sea. The reader may perhaps think that this point, the limited range of the upward and downward movements in post-glacial times, is being insisted on with wearisome iteration. But the insistence is necessary when we remember how constantly both geologists and naturalists, in order to account for anomalies in the geographical distribution of animals and plants, bring into play such movements. The argument is constantly used, that a certain species cannot cross the sea: therefore if it is found in an island, that island must once have been connected with the mainland. Nature is more full of resource than we imagine, and does not thus neglect her children. The cumulative effect of rare accidents spread over many thousand years is also far greater than may be thought by those who only consider what has been noted since means of dispersal have been studied scientifically. An examination of the south side of the Bristol Channel need not long delay us, except for two pieces of evidence which should not be passed over. In Somerset there are wide expanses of marsh land known as the Bridgwater and Glastonbury Levels. These greatly resemble the Fenland, and like it are underlain by a submerged rock-platform which has sunk in post-glacial times. But in this case we are able to fix a definite historical date by which all movement had ceased—it may have ceased much earlier, but we can prove that at any rate there has been no change of the sea-level subsequent to a certain date.