production of the Prismatic Structure of Basalt. 213 



ated masses, such as arragonite, hematite, wavellite, &c. But 

 observation proves that this is not the case; the taper prisms 



fi c . ig FI C .17. 



here, though in some respects simulating the forms of these 

 taper crystals, are not produced by crystalline forces, but by the 

 mechanical work of splitting up the cooling and contracting 

 mass. The principle of economy in work, which determines 

 the size as well as the form at starting, continues throughout the 

 entire production of the prism however deep it may enter into 

 the mass ; but if each prism continued to diminish in diameter 

 indefinitely in length, there would be an enormous waste of 

 work in splitting the deeper parts of the mass into prisms far 

 more numerous than would be necessary there to relieve the 

 contractile strain, which, when the temperature has become the 

 same is per unit length the same as it were at the surface. 

 Hence after the taper prisms have reached a certain length from 

 the surface, they are either found to end at a thin stratum of 

 irregular fragments and a fresh range of taper prisms to com- 

 mence (as in fig. 16 bis), whose diameters are the same as those 

 of the prisms at the surface of the mass, 

 or two or more of the taper prisms are fi c . /6 d™ 



fractured more or less transversely and ,^-r-"'"' 

 irregularly, and, ending there, start off as Va 

 a single prism of larger diameter and more \v ^ ~. : , 

 or less irregular in form (as seen in fig. 16), 

 all the prisms in the neighbourhood of 

 such changes becoming more or less irre- *0\mWf$f 



gular, either as to form or number of sides 



or both. Examples of both these modes of adjustment are 

 very frequent in Auvergne, the Cantal, and many other basaltic 



J 



