It is certain that old burrows collapse in the course of time; for as we shall see in the next chapter, the fine earth voided by worms, if spread out uniformly, would form in many places in the course of a year a layer 0.2 of an inch in thickness; so that at any rate this large amount is not deposited within the old unused burrows. If the burrows did not collapse, the whole ground would be first thickly riddled with holes to a depth of about ten inches, and in fifty years a hollow unsupported space, ten inches in depth, would be left. The holes left by the decay of successively formed roots of trees and plants must likewise collapse in the course of time. The burrows of worms run down perpendicularly or a little obliquely, and where the soil is at all argillaceous, there is no difficulty in believing that the walls would slowly flow or slide inwards during very wet weather. When, however, the soil is sandy or mingled with many small stones, it can hardly be viscous enough to flow inwards during even the wettest weather; but another agency may here come into play. After much rain the ground swells, and as it cannot expand laterally, the surface rises; during dry weather it sinks again. For instance, a large flat stone laid on the surface of a field sank 3.33 mm. whilst the weather was dry between May 9th and June 13th, and rose 1.91 mm, between September 7th and 19th of the same year, much rain having fallen during the latter part of this time. During frosts and thaws the movements were twice as great. These observations were made by my son Horace, who will hereafter publish an account of the movements of this stone during successive wet and dry seasons, and of the effects of its being undermined by worms. Now when the ground swells, if it be penetrated by cylindrical holes, such as worm-burrows, their walls will tend to yield and be pressed inwards; and the yielding will be greater in the deeper parts (supposing the whole to be equally moistened) from the greater weight of the superincumbent soil which has to be raised, than in the parts near the surface. When the ground dries, the walls will shrink a little and the burrows will be a little enlarged. Their enlargement, however, through the lateral contraction of the ground, will not be favoured, but rather opposed, by the weight of the superincumbent soil.