This paper may be regarded as the sequel to an earlier one (Rayleigh 1942) and deals with the formation of artificial pebbles under controlled conditions, and their comparison with the pebbles found in nature. As before, symmetrical pebbles having an approximate figure of revolution are chiefly considered. A series of marble pebbles, made experimentally by attrition of rectangular blocks by fragments of hard steel, is shown in comparison with natural pebbles. The series ranges from cylinders with rounded ends, to an approximately spherical figure, and then on to oblate forms ending in a disk with rounded edges. These are closely matched by a series of natural flint pebbles collected from the glacial gravel. Prolate or oblate spheroids, differing widely from the sphere, are not obtained in experiments of this kind, nor are they found in the gravel formation. An alternative way of making pebbles in the laboratory is by 'pothole' action. The experimental pothole is a cylindrical vessel containing water with a paddle revolving coaxially with it. The paddle maintains a vortex, which carries the stone round. Either the bottom or the wall of the vessel may be made of abrasive material. When the bottom is abrasive, the pebbles are of such a shape that they lie inside a spheroid of the same polar and equatorial diameter in contrast to the previous case, where they lie outside. When the sides are abrasive, the form tends to the spherical. This effect appears to depend on the well-known tendency of an elongated body to set itself athwart the stream; thus the end tends to rub against the abrasive wall. Spherical pebbles of considerable perfection can be made in this way. As regards concave pebbles, the discussion in the former paper is withdrawn. A new method of experiment is used, depending on the pitting of a small square of sheet glass, revolved in a box with the abrasive. If many broken flints are used, each comparable in size with the glass, the latter is chiefly pitted in the middle, tending to form a concavity. If a single flint only is used, the pittings on the polished surface are uniformly distributed, and there is no tendency to form a concavity. It would seem, therefore, that concavity is produced by the edges being more protected than the middle, the protection being given by pebbles other than the one which is making the wound at the given moment. This action is apart from the rounding of the edges, which ultimately spreads, invading the concavity and producing a general convexity as in ordinary pebbles.