GENERAL MEETING. 23 
other they will no longer curve to the same extent, and the slower 
one will curve more than the faster one. 
Now, instead of two single columns, there,are always two groups 
_ approaching one another, and these prevent the continuation of the 
columns beyond the bend, pinching them out and causing them to 
taper off as already observed in the quarry described. 
The difference in the systems of cracks of the lower and upper 
portion of this lava sheet may then be accounted for by a difference 
in the rate of cooling from the lower and upper surfaces, the more 
frequent fractures arising from the more rapid cooling, and the two 
systems proceeding from their initial planes until they blend in one 
another within the mass. 
If for any reason the cooling from one surface should take place 
irregularly and from any point proceed more rapidly than from 
others, it is evident that there would result a set of columns diverg- 
ing from this point as a focus. 
Besides the columnar fracturing, a division of the mass by trans- 
verse cracks, especially near the top of the lava sheet, is more or less 
noticeable. 
There remains to be considered the contraction exerted in all 
other directions through the mass. Since the uniform contraction 
of a homogeneous body acts equally in every direction through it, 
its effect corresponds to the equal shortening of the radii of a sphere 
of such a body. If through any resistance cracking or parting 
occurs it will take the form of concentric spherical shells. 
If for any reason the resistance to the contracting force in a par- 
ticular direction be counteracted by some other force acting in the 
same direction the parting will no longer be spherical, but ellipsoidal, 
as will be seen from Figure 4, where abc represents a section 
through the sphere along the radii of which contraction takes place. 
A uniform resistance in the direction of the radii represented by 
