DISCUSSION OF ISOSTASY 187 



apparent maximum value of which is 6.73 per cent*" of the original 

 volume. 



How far down voids can exist is not fully determine*!. Mr. Frank 

 Adams has subjected cylinders of granite with holes drilled in them to 

 pressures corresponding to a depth of 35 miles without completely closing 

 the apertures; hut, making allowance for time and heat, he limits his 

 conclusion to the statement that openings may be permanent at least as 

 far as .1 1 miles from the surface.*" Mr. P. W. Bridgman*^ has subjected 

 sealed hollow cylinders of glass to a pressure of 24,000 atmospheres with- 

 out measurable permanent distortion, but he feels no confidence that a 

 crystalline solid would behave in the same way as glass at a depth of 56 

 miles. He tells me that in his experiments on rock specimens provided 

 with drilled holes these closed not by plastic flow, but by crumbling of 

 the walls. Since Mr. Bridgman can command a pressure of 40,000 at- 

 mospheres, corresponding to a depth of over 90 miles, there is no doubt 

 that more information is in store for us. 



Pseudo-anomalies then certainly exist; indeed they seem of the order 

 of magnitude of the observed or apparent anomalies, namely, ± 0.017; 

 for, as was shown ahove, even tlie largest gravity anomaly observed in the 

 United States, 0.095, could bo accounted for by the presence of a batho- 

 lith less than 9 miles in diameter. But the gravity anomalies when con- 

 sidered with regard to sign have a mean value of 0.000, just as would be 

 the case were there only pseudo-anomalies and were isostatic compensa- 

 tion perfect at about 120 kilometers. It is barely possible that the real 

 anomalies are of the same order of magnitude as the pseudo-anomalies, 

 and that the mean value of each species with regard to sign is 0.000 ; but 

 were this the case I should expect greater local apparent anomalies than 

 Messrs. Hayford and Bowie have observed ; for there is little direct con- 

 nection between the pseudo-anomalies and tlie real ones, so that they 

 might be expected to reinforce one another at something like one-half of 

 the whole number of stations. But that the real anomalies and the 

 pseudo-anomalies should each average 0.000, though conceivable, is very 

 improbable. The oidy way I can see of reconciling the observations with 

 probability is to suppose that the real anomalies, though not zero, are so 

 small as compared with the pseudo-anomalies that their effect on the 

 average is insensible, or, in other words, that the real anomalies are small 

 quantities of the second order. 



This conclusion, if conceded, means that the earth below the level of 



<5.Tour. Wash. Acad. Sci., vol. 4. 1014, p. 420. 

 ^''.Toiir. Geol., vol. 20, 1012, p. 07. 

 « Phil. Mag., vol. 24, 1012, p. 63. 



XIV— BnLL. Geol. Soc. Am., Vol. 26, 1014 



