170 ON THE CONCRETIONARY AND 



in these cases, only a modification of the laminar 

 structure, produced by fissures. Where it occurs in 

 the sandstones, it appears to be more certainly refer- 

 able to this cause. When it was thought certain 

 that the spheroidal exfoliation of the cuboids of gra- 

 nite was a proof of a spheroidal concretionary struc- 

 ture, it was natural to consider these as prismatic 

 concretions. That will till be true should this be 

 proved; but as some of the cases of this nature are 

 unquestionably shown, in the chapter on the de- 

 struction of rocks, to arise from the action of the 

 atmosphere, the whole question must remain open for 

 further inquiry. 



The columnar structure, on account of its sym- 

 metry and artificial appearance, is unquestionably the 

 most interesting of all these modifications: an interest 

 not a little enhanced by the difficulty of explaining its 

 origin. It presents many varieties, and occurs in rocks 

 of very different characters. 



The most remarkable forms of this nature are those 

 which exist in the rocks of the trap family; in the 

 history of which will be found the names of the par- 

 ticular substances in which they occur, together with 

 such minuter details as are not necessary for the pre- 

 sent purpose. (Chap, xxxix.) These columns are of 

 various sizes, ranging from the diameter of even an 

 inch to one of many feet, and. in height, from a foot 

 to many hundred feet. They are almost invariably 

 associated in groups, so as to occupy the whole, or 

 portions, of the stratiform beds occasionally found in 

 the trap rocks. In these cases they are generally 

 parallel, with more or less of exactness; but, in some, 

 they are variously and irregularly implicated. Occa- 

 sionally, they are even intermixed with amorphous 

 matter of the same nature. They are, commonly, 



