64 
GILBERT. 
miles, and the nature of the crystallines beneath them is 
not known. The crystallines were assumed to have an 
average density of 2.70. The mean density of the sedi¬ 
mentary column at each station was subtracted from 2.70 
and the remainder multiplied by the thickness of the sedi- 
mentaries. The products were used in the subsequent com¬ 
putations, yielding the quantities of the column. 
Column 5 shows values under the theory of high rigidity 
(column 2) as modified by the geologic correction (column 4). 
Column 6 shows values under the theory of isostasy (column 
3) as modified by the geologic correction. 
The formation of the remaining columns depends for its 
justification on the postulate that the ideal result of the cor¬ 
rection of pendulum observations is uniformity. If all 
local conditions were duly accounted for, the resulting values 
would be identical for all stations. Therefore, in seeking to 
ascertain whether the use of a doubtful correction is ad¬ 
visable, it is proper to ask whether its application tends 
towards uniformity of result. To that end I have brought 
out the comparative discordance of the various sets of values 
by subtracting the mean of each set from the several values 
of the set and tabulating the remainders. 
These appear (for columns 2, 3, 5, and 6) in columns 2a, 
3a, 5a, and 6a. At the foot of each of these columns is given 
the mean of its numbers, regardless of sign, and these means 
are considered to be indices of the discordance of the sets. 
The mean of numbers in column 4a, derived from the geo¬ 
logic corrections, may be regarded as an expression of the 
average quantitative effect of the geologic corrections in 
modifying the inequalities of the gravity determinations. 
Putting the results of the comparison into words: 
(1.) The discordance, .064 dyne (column 2a), of values de¬ 
rived from the theory of high rigidity is reduced to .054 dyne 
(column 5a) by the application of the geologic correction. 
It is reduced by the entire quantitative effect of that correc¬ 
tion, .010 dyne (column 4a). 
(2.) The values derived under the theory of isostasy have 
