Page:Transactions of the Geological Society, 1st series, vol. 2.djvu/86

76 own island of Arran (that little abridgement of the world) nodules of spherical granite are found in the valleys which descend from Goatfield, decomposing on the surface in crusts, and marking decidedly the very construction which my supposition requires, in a much greater degree than is requisite for the purpose. Similar granite balls have been seen in other places, so that their existence is well ascertained, and this is one of numerous instances, where the decomposition of a rock gives us most useful information with regard to its structure, and where without that aid, we should never have divined the secret of its formation. It is certain that these balls, now rendered spherical by decomposition, have been quadrangular masses, and hence we may step, without any great hazard of unsound footing, to this general conclusion, that these masses of granite, which show marks of wearing on their surface with rapidities proportioned to their distance from a central point, have had their hardness, and probably their crystallization or formation, determined from that centre.

The analogy of this circumstance to the similar balls formed in basaltic rocks is illustrative of both the cases, and probably both will equally tend to confirm the opinions which have been held relating to the igneous origin of these substances. Thus, if for the sake of argument I may be allowed to assume that granite is of igneous origin, it will be easy to explain the peculiar appearances exhibited by that formation of granite, which, like those of  Cornwall and Arran and many others, is separated into cuboidal masses.

Here we must conceive, that in a homogeneous mass of fluid matter, crystallization had commenced from numerous centres at the same time. While there was yet space for the formation of successive solid deposits round any set of these imaginary centres, a spherical or spheroidal figure would be the result. As the surfaces