NOTES ON GRAVITY DETERMINATIONS. 
67 
for Its determination the discussion of an elaborate system 
of stations. Probably an inner circle should be given 
greater weight than outer rings, and some attention should 
be paid to geologic provinces. The corrections here used 
were determined from circular districts having a radius 
of thirty miles and without the use of weights.* Of the 
circles about Denver and Colorado Springs only those parts 
belonging to the Great plains were used. It was assumed 
that the very different history of the mountain region at the 
west barred it from use in the determination of the standard 
value of gravity. The corrections obtained were: Ithaca 
+ .006; Ellsworth, +.002; Colorado Springs, +.009; Denver, 
+.008. 
In the second column of Table II are the altitudes of the 
stations. The third column contains the correction in dynes 
for reduction to mean plain. In the fourth column are the 
values of gravity at the stations after correction for latitude, 
altitude, local topography, and height of mean plain. The 
fifth column gives the residuals after subtracting the mean of 
the eleven from the individual values. The average residual, 
.008 dyne, is a measure of the discordance of the results for 
the several stations. The residuals may also be used to 
determine the probable error, .002 dyne, of the mean grav¬ 
ity, 980.151 dynes, regarded as a standard, under the iso¬ 
static hypothesis, for the discussion of the remaining stations 
of the chain; and they can be used to determine the prob¬ 
able error, ± .007 dyne, of the value derived from a single 
station of the interior plain. It is to be noted that the 
probable error for the single station includes not only errors 
arising from the evaluation of the various corrections and 
errors of observation, but all local departures of the plain 
* When these paragraphs were written I had not seen Mr. Putnam’s 
manuscript. My “reduction to mean plain ” is his “ Faye’s reduction’’ 
(pp. 47, 52-55). His term is preferable, but I let my lines stand unchanged, 
because there is some interest in the fact that we independently selected 
the same method of discussion. Our numerical results differ chiefly be¬ 
cause he used a radius of 100 miles and I a radius of 30 miles in com¬ 
puting the correction. 
