16 L. A. Bane?' — Gravity Determinations at Sea. 



to sign is , 022 cm . The individual differences occasionally 

 amount to 015 1 " 1 . lie finally says : "All conclusions drawn 

 in the previous publications remain unaltered." These main 

 conclusions are : 



"The acceleration of gravity over the oceans traversed is 

 approximately normal and conforms with Ilelmert's gravity 

 formula of 1901. Pratt's hypothesis of isostatic adjustment of 

 the masses of the earth's crust is thus, except for local anoma- 

 lies, found to hold true generally. It can be regarded hence 

 as proved that the lesser density of the water of the oceans is 

 compensated for by the increased density of the masses below 

 the ocean bottoms." 



My contention is that this conclusion was already practically 

 embodied in Hecker's method of adjustment. The conclusion 

 may be true, but it can not be considered as proved by his 

 mode of attack. Since no attempt was made to test whether 

 another formula for normal gravity might not still better con- 

 form with the observations, the statement at the close of the 

 first sentence does not seem warranted. 



Observations on Ilecker's Ocean Gravity Work. 



1. No wholly satisfactory measure of the absolute accuracy 

 of the existing ocean gravity results can be secured by a mere 

 perusal of the publications. If an independent examination is 

 made and such checks applied as are possible, and when all 

 sources of error are considered, it will not be surprising if it be 

 found that many of the most recently published results are in 

 error by an amount approximating to 0*l om , or about 1/10,000 

 part of g. In view of the pioneer nature of the work, it would 

 have been desirable to have repeated observations, under dif- 

 ferent conditions, over all regions previously traversed. 



2. One of the chief sources of error is to be ascribed to in- 

 constancy of the corrections of the boiling-point thermometers 

 caused by their continued and protracted use ; the error thus 

 arising may at times transcend in importance all other ones, an 

 error in the temperature of - 01° C. corresponding to about 

 - 35 cm in g. Insufficient attention was paid to purely instru- 

 mental changes and corrections. Thus, for example, correc- 

 tions for the boiling-point thermometers of the Atlantic Ocean 

 w 7 ork of 1901 were used practically unaltered throughout the 

 subsequent cruises of 1901 and 1909 — after having once been 

 supplied by the Physikalische Reichsanstalt, the corrections 

 were never again redetermined. No separate examination of 

 the barometers by comparison with standard barometers appears 

 ever to have been made. The belief that such purely instru- 

 mental changes would be fully taken account of in the adjust- 



