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



upwards ; the second a like mass, with the bottom concave; and 

 the third a mass with the top surface convex and reposing upon 



s 







P 7> p 















r 

 S 



^ 

























r\\ \ 





















\ \~^ \ \ 





















fK \.A-' \ \ 



















*WKK\ \ A"\" 



















<x 

















F/C M A «l^ 



s 



w 





IT? 





""" 







'1 



FIQ.2ZB 



^^?PlW#fIiMWI 



a level plane, cooling taking place in each case from the top sur- 

 face only. The isothermal planes whose vertical sections are 

 represented in figs. 22, 22 A, 22 B by dotted lines will be seen to 

 diverge towards the shallower part of the mass in every case — just 

 as the ice which forms upon a pond, first forms and becomes 

 thickest at the shallower parts of the water, and thins out, it may 

 be to nothing, as it approaches the deeper parts — the under sur- 

 face of the ice (which at any instant is the isotherm of 32° Fahr.) 

 not being parallel to the level upper surface, but rising as it ap- 

 proaches the deeper and therefore still warmer portion of the 

 water, which, like our basalt, cools only from the surface. 



The sole necessary condition, therefore, to the production of 

 curved in place of straight prisms is sufficient inequality in 

 volume, and therefore in the rate of cooling, of the mass at dif- 

 ferent points of its surface, the measurements being taken nor- 

 mally to the cooling surface ; and curvature in the prisms must 



