slaty cleavage at a late stage in the deformation of the rock. These authors, it will be seen, have developed one of Darwin's principal propositions, as to the close connexion between rock-cleavage and foliation, and, in opposition to Sorby, consider the platiness of the original constituents to be of less importance. In support of their view, in regard to the late stage at which cleavage is induced, it may be noted that the crystals of pyrite and magnetite that sometimes occur in slates and in the allied foliated schists have developed at an earlier date as knots which oppose the cleavage or the foliation[52]. Darwin observed that mineral differences sometimes occur along bands parallel with the cleavage-planes. In such cases, the difference may be largely one of grain, shearing having broken down the minerals into a finer state along certain bands of movement[53]. Shearing of the rock may occur along any of the cleavage-planes, which are superinduced planes of weakness, and parts of the slate thus slide over others, just as the mineral flakes slide over one another in the directions in which expansion of the rock is possible. Where traces of the original stratification remain, it is easy to see if rock-shearing has occurred. Beds of different composition naturally take on cleavage in very different degrees. Sandy layers show the compression that has taken place by contorting; but they cleave very poorly, and in proportion to the amount of mud present in them. Where clayey and sandy layers alternate, and the direction of the cleavage is oblique to them, it is refracted, as it were, on passing from one layer to the other; it is more highly inclined to the bedding in the sandy layers and less so in the clayey layers. Hence a cleavage-surface forms a fold resembling the shape of an italic _S_ as it traverses each harder bed. Harker[54] and Leith[55] discuss the cause of this from somewhat different points of view. It is probable that such cleavage-planes as develop within the hard bed are approximately perpendicular to the direction in which the compressive force acts, because there is in such beds little possibility of lateral creep of the material along the bedding-planes. In the softer layers, we have to deal, not only with a tendency towards the rotation of platy particles until their flat surfaces are perpendicular to the direction of pressure, but also with a tendency of the same particles to flow along the bedding-planes. The resultant arrangement gives rise to a cleavage nearer to the bedding-planes than that in the more sandy layers.