392 Bev. 0. Fislier on the Theory of 



the same amount, the linear contraction being in that case 

 the same in all directions, it would become simply smaller, 

 without any tendency to either cracking or crumpling. 



2. If, however, the outer strata were to cool more than the 

 inner ones, it is clear that they would become too small to fit 

 the uncooled nucleus ; and this effect would reach down to 

 the level at which cooling, and therefore contraction, became 

 insensible. Supposing, then, that nothing further happened 

 to the rocks than a simple contraction according to Marriotte's 

 law, the strata must needs crack ; and we may imagine that 

 the crust would be divided up by fissures, widely gaping 

 towards the surface, into prisms similar in form to basaltic 

 columns, and reaching down to the uncooled matter. The 

 vertical thickness of a crust so cooled would be diminished 

 by the sum of the linear contractions of the thicknesses of 

 each infinitesimally thin shell in accordance with Marriotte's 

 law, and the circumference of each shell, not counting the 

 width of the cracks, would be shortened in proportion to the 

 entire fall of temperature which had been experienced by 

 that shell. The result in this second case, depending solely 

 upon Marriotte's law, would be independent of the time. 



3. Turning next to a third case, more nearly approaching 

 what might be supposed to occur to a solid earth during the 

 fall from the uniform high temperature of solidification to 

 that which is its present distribution, the rate of cooling 

 would not be the same at different depths. At any epoch, 

 since the surface assumed the constant temperature of the atmo- 

 sphere, the cooling at the surface is nil. At a certain depth, 

 where the cooling is insensible, it is again nil. At some 

 intermediate depth, depending on the time, the rate of cooling 

 is greatest ; and where it is greatest, there the rate of con- 

 traction will be greater than anywhere above or below that 

 depth. In the case we are considering it is not probable that 

 open cracks could anywhere be formed, unless just near the 

 surface, because the weight of the superincumbent matter 

 would press out the contracting shells laterally, so as to close 

 them up. Under these circumstances we could not in general 

 arrive at the change of dimensions by applying the coefficient 

 of linear contraction to the horizontal and vertical dimensions 

 separately of each shell ; but wherever the shell is extended 

 (or " stretched ") we can only apply the coefficient of volu- 

 minal contraction to the shell as a whole. 



Let us now fix our attention upon the condition of a par- 

 ticular shell of rock at the present epoch. We find it con- 

 tinuous and without open cracks, its temperature is falling, 

 and, owing to the contraction of the sphere of matter interior 



