lowest land-surface, and the nature of their enclosed fossils, suggest long-continued but intermittent subsidence; I can see, however, no indication of a reversal of the process. The land-surface is carried beneath the water, the estuary then silts up, becomes fresh water, marsh-plants grow, and even trees may flourish on this marsh before it subsides again. But there is no sign that the strata were ever raised above the level to which ordinary floods could build up an alluvial flat. The land-surfaces seem always to have been swampy, and bed succeeds bed in fairly regular sequence, without the deep channelling we might expect to find when an alluvial flat was raised to a noticeable extent above the level of high water.
The width of the Bristol Channel makes it clear that this gulf must occupy a submerged valley of great antiquity. It becomes therefore of interest to enquire whether the wide valley is correspondingly deep, or whether its rocky floor is found at the same shallow depth as in the case of the other river-valleys which we have been considering. The wide valley may have been formed in either of two ways. It may have been excavated as a deep valley with its bottom many hundred feet below the present sea-level. Or it may have commenced as the shallow valley of a big river with exceptionally powerful tides, and as this river swung from side to side it greatly widened its valley without making it any deeper.
