212 Mr. R. Mallet on the Origin and Mechanism of 



single sheet, but is broken up into partial waves with divergent 

 and, as may be said, accidental wave-paths, so that fracture by 

 contraction takes place in many different directions. The strata 

 of irregular fragments thus produced are not invariably found in 

 nature, being sometimes replaced by thin beds of clays or ochres 

 whose incoherence and small conductivity admit of more regular 

 cooling at the extremity of the prisms. Such irregular frag- 

 ments are also found occasionally converted into ochre by de- 

 composition. If the thickness of a tabular mass varies but 

 slightly and gradually, so that the section is slightly wedge- 

 shaped, and cooling takes place mainly from the top surface, 

 then the prisms may be straight and perpendicular to the 

 top surface throughout the mass, and only separated from the 

 bottom by a stratum of irregular fragments. But if (as in 

 fig. 15) the bottom of the mass should be more highly inclined 

 and cooling take place both from 

 the top and the bottom, then will 

 there be two ranges of straight 

 prisms formed, perpendicular re- 

 spectively to the top and to 

 the bottom surfaces, and separated 

 by an intermediate stratum of ir- 

 regular fragments — the respective 



lengths of the two ranges at any point depending on the rela- 

 tive rates of cooling, as already stated with reference to fig. 13 

 and others preceding it. 



The directions of the axes of the prisms being dependent 

 upon the contour passing through all the points of actual spirt- 

 ing at any instant (that is, upon the bounding contour of the ori- 

 ginal mass, as already fully explained), it is obvious that a mass 

 presenting a curvilinear contour of cooling must produce diver- 

 gent or convergent ranges of prisms, according as the cooling 

 surfaces are convex or concave, the separating points of the 

 prisms being every where normal to tangent planes passing 

 through them at the surface of the mass (as in figs. 16, 17, and 

 and 18) *. Convergent prisms must therefore be taper prisms ; 

 and it might be supposed that this tapering might continue 

 until they all converge to a point, as do the crystals of 

 certain minerals occurring in botryoidal or reniform and radi- 



* The isolated aggregations of prisms leaning against each other and all 

 slopiug towards the centre (like hop-poles stacked in water), noticed by 

 Mr. Scrope at several places in Auvergne, are the remaining or central 

 portions of prismatic masses (as in the lower portions of figs. 18 & 19, and 

 shown in Mr. Scrope's fig. 23, 'Volcanoes,' second edit. p. ( J4), all the other 

 portions of the mass having been removed by denudation, as intimated by 

 that author. 



