Bar rell — Growth of Knowledge of Earth Structure. 163 



"The argument for the contractional hypothesis presupposes 

 that the earth-mass may be considered as consisting 1 of two por- 

 tions, a cooled exterior of undetermined (though probably com- 

 paratively' small) depth, inclosing a hot nucleus. . . . The 

 secular loss of heat, it is assumed, would be greater from the 

 hot nucleus than from the exterior, and the greater consequent 

 contraction of the nucleus would therefore gradually withdraw 

 the support of the exterior, which would collapse. The result- 

 ing strains upon the exterior would be mainly tangential. 

 Owing to considerable inequalities in the ability of different por- 

 tions to resist the strains thus developed, the yielding would take 

 place at the lines, or regions of least resistance, and the effects 

 of the yielding would be manifested chiefly, or wholly, at those 

 places, in the form of mountain chains, or belts of table-lands, 

 and in the disturbances of stratification. The primary division 

 of the surface into areas of land and water are attributed to the 

 assumed smaller conductivity of materials underlying the land, 

 which have been left behind in the general convergence of the 

 surface toward the center. Regarding these as the main and 

 underlying premises of the contractional argument, it is con- 

 sidered unnecessary to state the various subsidiary propositions 

 which have been advanced to explain the determination of this 

 action to particular phenomena, since the main proposition upon 

 which they are based is considered untenable. 



There can be no reasonable doubt that the earth-mass consists 

 of a cooled exterior inclosing a hot nucleus, and a necessary 

 corollary to this must be secular cooling, probably accompanied 

 by contraction of the cooling portions. But when we apply the 

 known laws of thermal physics to ascertain the rate of this 

 cooling, and its distribution through the mass, the objectionable 

 character of the contractional hypothesis becomes obvious. 



That Fourier's theorem, under the general conditions given, 

 expresses the normal law of cooling, is admitted by all mathe- 

 maticians who have examined it. The only ground of contro- 

 versy must- be upon the values to be assigned to the constants. 

 But there seem to be no values consistent with probability which 

 can be of help to the contractional hypothesis. The applica- 

 tion of the theorem shows that below 200 or 300 miles the cool- 

 ing has, up to the present time, been extremely little. . . . 

 At present, however, the unavoidable deduction from this 

 theorem is that the greatest possible contraction due to secular 

 cooling is insufficient in amount to account for the phenomena 

 attributed to it by the contractional hypothesis. 



The determination of plications to particular localities pre- 

 sents difficulties in the way of the contractional hypothesis which 

 have been underrated. It has been assumed that if a contraction 

 of the interior were to occur, the yielding of the outer crust 

 would take place at localities of least resistance. But this could 



