346 THE AGE OF THE EARTH. 



rock under high pressure are of conrse beyond the reacli of experimen- 

 tation, hence we are forced to h)ok to those of the avaihible nmterials. 

 Isometrics from high-pressure observations have been found to slope as 

 foHows: 



Atniospbcre per <^C 



Ether -. 8-7 



Alcohol 10-5 



Thymol 13-9 



Dyphonylamiue 15 '4 



Paratolnidiiie 13-9 



Glass, coinpiitetl 10 "0 



Since ether boils at 34° C. and dyphenylamine at SIO'^ C, the range 

 here given is wide. It is reasonable, therefore, to take the mean value, 

 lli'5 atmospheres per ° C, as an index of the slope sought for. In the 

 Kelvin earth this rate occurs between O-OIO and 0-015 of radius, the 

 crust being 0-0065 of radius thick. In so far therefore as the isomet- 

 rics can be regarded as parallel straight lines with a slope of the order 

 of the value given above, convection can only have taken place in the 

 first 52 miles of the initial couche of fusion, and in the present resid- 

 ual couche of 200 miles only the upper 2C> miles would be subject to 

 convection. In younger earths the above value per ° 0. will be found 

 much nearer the surface, so that in them convection would be confined 

 to a shell, which is shallow ni proportion as the earth is young. Ini 

 tially when the whole earth was at one temperature there could have 

 been no convection, since the change of temperature in depth was w«7, 

 but the change of density due to pressure was always pronounced. In 

 the case of the 1,7410 C. earth the zone of convection would have early 

 been covered and extinguished by the thickening crust, and therefore 

 would have played no very important part in accelerating the loss of 

 heat, and thus for this particular initial excess is of small effect in 

 shortening the estimate of earth's age. 



h. The direct effects of heat and pressure upon the conductivity of 

 matter under such high temperature and pressures are also beyond lab- 

 oratory investigation, and again we are driven to use the determined 

 conduci^ivity value unmodified, or seek for some other property which 

 may be considersd as its approximate measure. Such an index is found 

 in viscosity, which if not of high quantitative signilicance in defining 

 the changing values of terrestrial conductivity in depth, nevertheless 

 aflbrds data applicable at least for determining the sign of an import- 

 ant coirection. 



Dr. I>arus has lately determined that at least 200 atmospheres of 

 pressure are required per 1° C. in order that viscosity may remain con- 

 stant. Examining several temperature distributions of the chart and 

 applying the computed augmentation of earth pressure, itai)pears that 

 the required relation (200 atmospheres to 1° C.) is found at successively 

 lower depths for successively higher values of initial excess and age- 

 In the 1,741° C. case the relation after 20x 10 « years' cooling is found 

 at about 010 of radius counting from the surface, where the vertical 



