on this question. Thus we may possibly find that this disputable appearance forms part of a series of very common phenomena, and that although it may possess a certain general resemblance to the effects of mechanical action, it is, in fact, produced by chemical agencies.
There are two circumstances to be regarded in describing the fracture and texture of a rock, its small or artificial fracture and microscopic texture, and its large or natural fracture and texture, as they are determined and exhibited by the effects of time or decomposition.
The instances which might be adduced to illustrate these differences are sufficiently familiar. In the case particularly of the large disposition and fracture of rocks, a striking difference occurs in the often approximated beds of trap and sandstone. Both form parallel beds, and both on the application of force break into nearly similar fragments, yet the natural division of the sandstone bed shows an horizontal tendency, or one parallel to the plane of the bed, while that of the trap is vertical to it. From the vertical fracture a series of gradations occurs which at length assumes the perfectly geometrical form of polygonal columns. In this case then, we have a form still more perfectly mechanical and regular than that of the most even stratification, and produced by a species of crystallization, a tendency to decided forms in those particular rocks, with the laws and causes of which we are at least as well acquainted as we are with the laws that determine the figure of a quartz crystal. At present they are both equally inexplicable. There is no further difficulty in conceiving that a rock may constitute a huge bed separable into horizontal laminæ as regular as the strata of a mechanical deposit, than in conceiving that the island of Staffa is separable into columnar fragments, or the rock of Devar into vertical laminæ. It is true that we have not yet produced any instance of continuous, horizontal laminar concretions, which are incontrovertibly not mechanical, yet