174 ON THE CONCRETIONARY AND 



produce. The following facts seem to prove that 

 the columnar structure was the consequence of the 

 same action. 



In Rum, the columnar sandstone actually lies 

 -beneath a mass of trap ; so that the fact of their 

 simultaneous presence, at least, is proved. This, 

 it is true, is as yet a solitary instance ; hut here, 

 fortunately, direct experiment comes in aid of the 

 supposition that the action of heat has produced the 

 columnar structure of sandstone. The sandstones 

 used for the hearths of iron-furnaces, after long ex- 

 posure to the heat of these, become divided into 

 polygonal prisms, exactly resembling those of the 

 natural prismatic sandstones, on a small scale. In 

 this case, there is no shrinking, as in dried clay, to 

 account for the appearance ; the sides remaining in 

 perfect contact, just as they are found to do in the 

 columnar traps. The same circumstance occurs in 

 those sandstones which are heated for the purpose of 

 making roads in Derbyshire. Here therefore it is 

 directly proved, that heat is capable of inducing the 

 prismatic structure in a solid sandstone ; and, that 

 this is not the developement of an original concre- 

 tionary structure, is proved by the fact, that in the 

 hearthstones which have undergone this change, the 

 arrangements of the prisms is always vertical to the 

 plane of the stone ; a remarkable analogy to their 

 mode of arrangement in the trap rocks. 



Ignorant as we are of the nature of the concre- 

 tionary structure, it is still certain that it bears a kind 

 of analogy to crystallization ; and experiment proves 

 that this arrangement, if it be not rather a concre- 

 tionary one, may take place in rocks, without flui- 

 dity. It is also remarked in the chapter on curva- 

 tures, that a curved structure is sometimes developed 



