The pot-holes and “cirques” in calcareous rocks with no outlet at the surface, may also be accounted for by the operation of the same causes as those which have produced caves. Each represents the weak point towards which the rainfall has converged, caused very generally by the intersection of the joints. This has gradually been widened out, because the upper portions of the rock would be the first to seize the atoms of carbonic acid, and thus be dissolved more quickly than the lower portions. Hence the funnel shape which they generally assume, and which can be studied equally in the compact limestone or in the soft upper chalk. They are to be seen on a small scale also in all limestone “pavements.” Sometimes, however, the first chance which the upper portions of the funnels have of being eroded by the acidulated water, is more than counter-balanced by the increased quantity converging at the bottom, and the funnel ends in a vertical shaft. If the area in the rock thus excavated be sufficiently large to allow of the development of a current of water, the mechanical action of the fragments swept along its course will have an important share in the work, as we have seen to be the case in Helln Pot. _Caves not generally found in Line of Faults._ In some few cases the lines of weakness which have been worn into caves, pot-holes, ravines, and valleys, may have been produced, as M. Desnoyers believes, by subterranean movements of elevation and depression; but in all those which I have investigated the faults do not determine the direction of the caverns. The mountain limestone of Castleton, in Derbyshire, offers an example of caves intersecting faults without any definite relation being traceable between them. The ramifications of the Peak cavern traverse the Speedwell Mine nearly at right angles, and the water flowing through it has been traced, Mr. Pennington informs me, to a swallow-hole near Chapel-en-le-Frith, running across two, if not three faults, which are laid down in the geological map. As a general rule caverns are as little affected by disturbance of the rock as ravines and valleys which have been formed in the main irrespective of the lines of fault.