THE AGE OF THE EARTH. 345 



from top to bottom), nucl since it emerges on the minimum i-iite it lias 

 already iallen below the admissible tangent. 



Gradient d, 1,050° C. and 15x10*' years, just cooled to the maximum 

 surface rate, has still an inadmissible fluid shell, but if refri.yeration 

 had beeu continued for 7 x 10'' to 9 x lO*' years more the line would have 

 fallen below the solidity line and its surface rate would not have passed 

 tlie mean value. Hence a 1,050'^ 24x10'' year earth is possible and 

 marks about the superior limit admissible for initial excess. 



From the iwint of view of age no greater time of cooling is allowable 

 than enough to bring the gradient for any initial excess to the mean 

 surface rate. Thus the condition for excess and age exclude a line of 

 over 2,000° C and 24 x 10'^ years. Conductivity remaining of the value 

 used, any higher excess involves fluidity, and anj^ greater age an inad- 

 missible surface rate. 



To the extent, therefore, that solidity is a valid criterion and so far 

 as the melting temperature of diabase maybe supposed to apply to the 

 depth examined, there is no escape from an earth of the low age and 

 temperature given except by impugning the rate of surface augmenta- 

 tion and the value of rock conductivity here employed. 



Whoever has examined the B. A. committee's reports and summaries 

 on underground temperatures must have realized the obstacles to the 

 evaluation of a true mean rate. The range of observations is wide, 

 from high rates due to residual vulcanism to low ones produced by 

 neighboring bodies of cold water, such as are described by Wheeler 

 from mines near Lake Superior.* It is not however likely that by 

 rejecting anomalies and assigning probable weight to further observa- 

 tions the i^resent value will be moved to an important extent. 



We have seen that all probable distributions of earth temperature 

 involve in the initial stage a great solid nucleus, practically the whole 

 body of the earth, with a shallow surface shell of fusion. In the case 

 of the 1,741° C, 20 x 10" year earth there was an initial melted shell 

 of 53 miles. Obviously it can not be correct to apply the rock-conduc- 

 tivity value obtained at air temperatures and normal pressure to even 

 so young and cool an earth with its couche of an initial temperature 

 of 1,741° C. and a pressure difference between the top and bottom 

 of 22,000 atmospheres. The probable method of cooling the couche into 

 solidity, involves three corrections of the accepted rate of refrigeration : 

 a, acceleration of the process by i)Ossible convection; h, the direct 

 effect of heat and j)ressure up;)n conductivity, and c, the relative con- 

 ductivity of matter at the some temperature, liquid and solid. 



a. Convection. Leaving out of present consideration possible poly- 

 merization of the magma, or the descent of solid bodies of crust, ver- 

 tical transfers of the liquid matter in the fused couche depend upon 

 differences of density and this upon the ratio of the rates at which 

 density is raised by pressure and lowered by heat. Isometrics of melted 



Am. Jour. ScL, 1886, vol. xxxii. 



