1307 
differences in the radius of curvature, or in the values ofy, have a 
large influence on the compression, and it seems not impossible that 
the resulting value of ¢ has been influenced by inaccuracies in the 
reductions. Discussing the large difference between the compressions 
found by Bresse, (et = 299.15) and Crarke (293.47) partly from 
the same observations, HerMerr *) asserts that this difference can be 
fully explained by a difference of a few meters in the adopted 
height of the geoid over the normal surface. If this is so, we can 
expect that considerably larger differences of the isostatic reduction 
will lead to similar effects *). 
For these reasons it appears to me that the agreement of the 
three values (I), (II) and (II) ean only be accidental. It is not at 
all certain a priori whether they refer to the same normal surface, 
and their uncertainty undoubtedly is considerably larger than would 
be inferred from the mean errors. *) 
From the lunar parallax we found in the preceding paper 
er OOB Be see ee od ol et 7 QLD 
We also showed that the value 296.0 cannot be said fo be 
excluded by the observations. 
The lunar theory gives J, from which e is found by the equation 
(1). The principal term, which is commonly used for the deter- 
1) Geoid und Erdellipsoid, Le. p. 18. 
2) The values of « derived from the American determinations by different methods 
of reduction (and different combinations of stations) are widely divergent. Thus 
e.g. from the observations in the United States and in Alaska by the isostatic 
method 300.4 +0.7 and by the free air method 291.2 + 0.7. See Bowie, lc. p. 26. 
The former of these should properly be quoted instead of (II’) as the final result 
from the American determinations. 
5) Hetmert’s formula of 1901, from which (IL) is derived, reduced to the Pots- 
dam system, is 
g = 9.78030 [1 + 0.005302 sin® p — 0.000007 sin? 2 p]. . (a) 
With the compression ¢—!= 296.0, and a constant correction of + 0.00011 
this becomes 
g = 9.78041 [ 1 + 0.0052764 sin? p — 0.0000074 sin? 2g]. (B) 
The residuals of these two formulas for different zones of latitude are as follows, 
expressed in units of 0.00001 : 
Zone 5° dae 25° 35° 45° 55° 65° 75° 
(a) a 0 — 20 +6 + 6 i —7 — 3 
(A) ie In Si eee eee) Te "| ede haa 
The m.e. of each of these residuals is +11. The residuals # naturally are 
somewhat systematic, but they are not larger than (z\, and can very well be due 
to errors of observation or inaccuracies in the reductions. A new discussion on 
the basis of the theory of isostasy, and including the valuable material, which has 
become available since 1900, is very desirable. [Note added in the English translation). 
