Ch. VI.] PRIMARY ROCKS. 95 



All these concretionary forms of the granitic rocks, deve- 

 loped by a partial decomposition, suffer a more rapid dis- 

 integration at their solid angles, by which the latter are 

 gradually rounded off, and the blocks are converted into 

 spheroidal masses, varying, of course, according to the figure 

 of the original concretions ; these are, for the most part, more 

 or less oblate or flattened, though they may occasionally be 

 seen as spheres, so regular, indeed, that they appear to have 

 owed their form to the chisel, and remind one of the granitic 

 shot fired by the Turks at our ships in the Dardanelles. 



This universal reduction of the granitic concretions to the 

 spheroidal form, by the action of the elements, a property 

 which is also common to greenstone, basalt, and other horn- 

 blende and crystalline rocks, is a fact which calls for farther 

 investigation. 



Dr. Macculloch is of opinion, that this spheroidal structure 

 is original, and that the cuboidal masses of these rocks are 

 subordinate thereto, being the result of the interference of 

 the contiguous spheres with each other. He conceives " that 

 in a homogeneous mass of fluid matter, crystallization had 

 commenced from numerous centres at the same time; and 

 that, 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 of these spheroids approached each other, the suc- 

 cessive crusts would interfere, and the remaining intervals 

 would be filled by portions of spheroidal crusts, until the 

 cuboidal figures of all the contiguous masses were completed; 

 thus forming that aggregated mass of cuboids which we 

 witness in the granites of this aspect which remain uninjured 

 in their places. We need not be surprised that this regu- 

 larity is not more constant, nor the forms more perfect, as we 

 are unacquainted with the numerous circumstances which 

 may determine the several centres of crystallization, or which 

 may interfere with the ultimate regularity of the resulting 



* Geol. Trans., vol. ii. p. 76. 



