232 



SCIENCE 



[N. S. Vol. XXVII. No. 684 



serve as a guide are those in massive rocks, 

 especially the nearly anhydrous basaltic, an- 

 desitic and rhyolitic massives. All cases 

 where there are local evidences of heat due to 

 thermal springs, etc., should be excluded, and 

 ■when for a normal rock the gradient is un- 

 usually high, it should be considered as sus- 

 picious. In short, for the present problem 

 the lower gradients in massive rocks are those 

 most likely to give a correct value of the 

 earth's age. So far as I can judge, the 

 gradient of 1° F. in 77 feet is not much, if at 

 all, too low from this point of view. 



The accompanying diagram represents the 

 temperature excess curve for the 60-million- 

 year earth as a full line. It is asymptotic to 

 the line V'a and involves no tidal instability. 

 A dotted curve in the diagram shows the tem- 

 perature of the same earth when 30 million 

 years of age. At that time the level of easiest 

 fusion, or the eutectic level, was much nearer 

 the surface than 140 kilometers in fact at 

 about 86 kilometers, and the increment of 

 temperature needful to produce fusion at the 

 eutectic level was smaller, only some 80° in- 

 stead of 140° C. At no stage of the life of 

 this earth was there tidal instability. Only 



in the earliest stages did the curve cross the 

 diabase line Dh, and that only at less than 

 40 miles from the surface, where by hypoth- 

 esis the rocks are andesites or rhyolites and 

 less fusible than diabase. On the other hand, 

 the temperature of the globe at great depths 

 is high, 2,000° being reached within a hundred 

 miles of the surface. 



Perhaps the least satisfactory of the as- 

 sumptions made in this discussion is that the 

 layer of rocks less fusible than diabase is 40 

 miles in thickness. To obtain an idea of the 

 importance of an error in this assumption I 

 have computed the gradient, assuming the re- 

 fractory layer to be only 30 miles thick in 

 the 60 X 10° year earth, the other data re- 

 maining unchanged. This calculation gives 

 1° r. in 79.2 feet, so that the effect of even a 

 very large error in estimating the thickness 

 of the refractory layer is not great. 



In the course of time it should be possible 

 to obtain better values of the constants than 

 I have employed. I urge a careful revision 

 of surface temperature gradients in the sense 

 of the remarks in a preceding paragraph, 

 accurate determinations of the temperature of 

 lava as it flows from the vents, and above all 



