264 CARNEGIE INSTITUTION 



and distribution of internal heat. In this case, manifestly, the 

 greatest heat would arise where there was the most compression, 

 and the distribution of internal temperature would have a genetic 

 relation to the distribution of internal pressures. As the second 

 hypothesis also makes pressure a genetic factor in determining the 

 temperature of solidification, these two hypotheses have much in 

 common, so far as thermal distribution is concerned. In their geo- 

 logical bearings they differ somewhat radically from the heretofore 

 more prevalent hypothesis of uniform initial temperature ; for, if 

 the heat gradient rises continuously — though not uniformlj- — to the 

 center, there must be an outward flow of heat at all points, and hence 

 the w^hole body of the earth must participate in the thermal changes, 

 and in the effects of these on earth movement and deformation. 



In these preliminarj- studies it was observed that the nature of the 

 pressure curve derived from the Laplacian law of density, or from 

 any other probable law, was such as to suggest that there would 

 ensue an internal redistribution of heat of such a nature that the 

 outer portions, neglecting the surface shell, might experience a rise 

 of temperature independently of any surface loss of heat. 



These considerations, briefly outlined here to render more intelli- 

 gible our working scheme, are so fundamental that it has seemed 

 worth while to spend some time in definitely developing the three 

 hypotheses into such specific forms as would secure from them their 

 best working stimulus and their greatest suggestiveness. This I 

 have attempted to do. 



The origin and distribution of internal heat being thus funda- 

 mental and critical, Mr. A. C. I^unn has undertaken a mathematical 

 inquiry into the thermal effects of compression under the accretion 

 hypothesis. As the data relative to compressibility, conductivity, 

 specific heat, and other factors involved cover only the small range 

 of conditions available in experimentation, and as even the possi- 

 bilities of experimental determination have been as yet by no means 

 exhausted, either in range or in accuracy, the method of multiple 

 hypotheses has been used to cover the probable range of variation 

 that may arise under interior conditions. The work is also being 

 formulated so that further extensions of application may be easily 

 made if required. AVhile it is to be regretted that laboratory data 

 are not better, there is compensation in the fact that such an inquiry 

 as is now in hand may show in what lines the tedious and expensive 

 work of experimental determination may be directed with the great- 

 est prospect of fruitful results and what lines may be avoided as 



