By the use of these figures, the following values are obtained for " g " : 



No. 5. No. 7. No. 21. 



Christchurch, 1910 .. (980-518) 980-528 980-533 cms. per sec, per sec. 



1913 .. (980-571) (980-554) (980-542) 



The mosb probable value I think is 980-534 (which is about 0-022 more than that 

 obtained by Bernacchi*), with a probable uncertainty of 0-010. 



It was at first the intention to compare the observed value of " g " with the value 

 calculated from Helmert's formula, corrected for height above sea-level, for density 

 of the earth's crust at the observing station, and for the attraction of neighbouring 

 land masses, but this course appears undesirable in a report of this nature. 



Hayford and other workers of the U.S. Coast and Geodetic Survey have shown 

 that there is, in the United States, a tendency to isostatic compensation, and have 

 derived figures for the most probable depth of compensation in the U.S., which varies 

 according to the assumption as to the relation between compensation and depth, the 

 data being insufficient to decide between different assumptions. This tendency towards 

 isostatic compensation has been confirmed in other parts of the world, and it is clear 

 that a closer approximation to compensation would be obtained if the further assump- 

 tion were made that the depth of compensation varied from place to place. 



The principle of isostasy can thus not be substantiated, though clearly a tendency 

 to isostatic compensation exists. It appears to the writer that useful results can only 

 be obtained from the application of the principle when large areas of the earth's surface 

 are considered, the anomalies then outstanding at different places being most probably 

 evidence of a tendency to uplift the earth's crust or to depress it. Certainly, if a ten- 

 dency to isostatic compensation exists, but it is not complete, the values of " g " at 

 any place must be dependent somewhat on the intensity of erosion at that place, or, 

 in the Antarctic, largely on the intensity of glacierisation and the rate at which the 

 ice masses on the continent increase or decrease. 



The acceptance of this principle leads inevitably to question to what extent the 

 ratio of " g " at two stations on the earth's surface will remain constant and inde- 

 pendent of time, and it would certainly appear that the possibility of a measurable 

 variation, even in quite short periods of time, should receive consideration. 



October Wk, 1921. 



* National Antarctic Expedition, 1001-1904 Physical Observations. 



91 



