Geologists, however, sometimes speak of the submerged forests as owing their present position to various accidental causes. Landslips, compression of the underlying strata, or the removal of some protecting shingle-beach or chain of sand-dunes are all called into play, in order to avoid the conclusion that the sea-level has in truth changed so recently. The causes above mentioned have undoubtedly all of them affected certain localities, and it behoves us to be extremely careful not to be misled. Landslips cannot happen without causing some disturbance, and a careful examination commonly shows no sign of disturbance, the roots descending unbroken into the rock below. It is also evident in most cases that no landslip is possible, for the “forest” occupies a large area and lies nearly level. Compression of the underlying strata, and consequent sinking of the land-surface above, is however a more difficult matter to deal with. Such compression undoubtedly takes place, and some of the appearances of subsidence since the Roman invasion are really cases of this sort. Where the trees of the submerged forest can be seen rooted into hard rock, or into firm undisturbed strata of ancient date, there can, however, be no question that their position below sea-level is due to subsidence of the land or to a rise of the sea, and not to compression. But in certain cases it is found that our submerged land-surface rests on a considerable thickness of soft alluvial strata, consisting of alternate beds of silt and vegetable matter. Here it is perfectly obvious that in course of time the vegetable matter will decay, and the silt will pack more closely, thus causing the land-surface above slowly to sink. Subsidence of this character is well known in the Fenland and in Holland, and we must be careful not to be misled by it into thinking that a change of sea-level has happened within the last few centuries. The sinking of the Fenland due to this cause amounts to several feet. The third cause of uncertainty above mentioned, destruction of some bank which formerly protected the forest, needs a few words. It is a real difficulty in some cases, and is very liable to mislead the archaeologist. We shall see, however, that it can apply only to a very limited range of level.