We thus see that many worms lived beneath the floor and the walls of the atrium at the time when the excavations were made; and that they afterwards almost daily brought up earth to the surface from a considerable depth. There is not the slightest reason to doubt that worms have acted in this manner ever since the period when the concrete was sufficiently decayed to allow them to penetrate it; and even before that period they would have lived beneath the floor, as soon as it became pervious to rain, so that the soil beneath was kept damp. The floor and the walls must therefore have been continually undermined; and fine earth must have been heaped on them during many centuries, perhaps for a thousand years. If the burrows beneath the floor and walls, which it is probable were formerly as numerous as they now are, had not collapsed in the course of time in the manner formerly explained, the underlying earth would have been riddled with passages like a sponge; and as this was not the case, we may feel sure that they have collapsed. The inevitable result of such collapsing during successive centuries, will have been the slow subsidence of the floor and of the walls, and their burial beneath the accumulated worm-castings. The subsidence of a floor, whilst it still remains nearly horizontal, may at first appear improbable; but the case presents no more real difficulty than that of loose objects strewed on the surface of a field, which, as we have seen, become buried several inches beneath the surface in the course of a few years, though still forming a horizontal layer parallel to the surface. The burial of the paved and level path on my lawn, which took place under my own observation, is an analogous case. Even those parts of the concrete floor which the worms could not penetrate would almost certainly have been undermined, and would have sunk, like the great stones at Leith Hill Place and Stonehenge, for the soil would have been damp beneath them. But the rate of sinking of the different parts would not have been quite equal, and the floor was not quite level. The foundations of the boundary walls lie, as shown in the section, at a very small depth beneath the surface; they would therefore have tended to subside at nearly the same rate as the floor. But this would not have occurred if the foundations had been deep, as in the case of some other Roman ruins presently to be described.