66 DONALD C. BARTON 
be +0.001 cm. sec.~* In the surveys in the Ries area of Bavaria, the 
differences between the determination of relative gravity by the 
torsion balance and the pendulum for six different lines were 0.003, 
0.000, 0.000, 0.002, 0.003, 0.001 cm. sec.~* The lines ranged from 4 to 
6 kilometers in length.'! In the Ural Emba district of Russia, the 
differences were 0.001, 0.002, 0.007, 0.004, 0.003, cm. sec. * The length 
of the traverses ranged from 9.5 to 15 kilometers.? In the Grosny dis- 
trict of Russia, the differences were 0.001, 0.002, cm. sec.~? and the 
lengths of the traverses ranged from 9.5 to 19 kilometers. As the prob- 
able error of the pendulum observations was +o0.001 to +0.002, and 
as errors of three times the probable error should be expected to be 
fairly common, the torsion balance surveys gave as accurate a de- 
termination of relative gravity as did the pendulum observations. 
But from the data, it is impossible to tell whether the determinations 
of relative gravity by those torsion balance surveys were more ac- 
curate than those by the pendulum. 
The accuracy of the torsion balance determinations of the gradient 
has been checked by the writer by calculation of the normal north- 
ward (planetary) gradient from torsion balance observations and com- 
parison of that value with the theoretical value. The theoretical value 
was 7.14 — north in comparison with the value of 7.4 E., N. 34°E., 
which was calculated from the torsion balance observations.® 
The present paper reports the determination of the effective ac- 
curacy of the (horizontal) gradient which is observed with the Eotvés 
torsion balance and of the accuracy of relative gravity which is cal- 
culated from the gradient. 
The data on which the study is based consist of the error of closure 
in relative gravity in forty-five closed traverses from commercial 
torsion balance surveys. The length of the traverses ranged from 3 
to 144 kilometers and in twenty-eight of the traverses, the lengths 
were evenly distributed over the range of 10 to 4o kilometers. The 
interval between stations on thirty-two of the traverses was 300-500 
meters, on five traverses, 150+ meters, and on eight traverses, 600— 
800 meters. The instruments which were used on the surveys mainly 
’ Karl Jung, ““Drehwagemessungen im Ries bei Nordlingen,” Zeitschrift fiir Geo- 
physik, Vol. 7 (1931), pp. 18-10. 
2 B. Numerov, “‘Results of the General Gravity Survey in the Emba District,” 
Zeitschraft fiir Geophysik, Vol. 5 (1929), pp. 268-70; and ‘Results of Gravitational Ob- 
servations in the Region of Grosny in 1928,” ibid., Vol. 5 (1929), pp. 271-75. 
§ Donald C. Barton, ‘Gravity Measurements with the Eétvés Torsion Balance,” 
in “Physics of the Earth,’’ Vol. II, “The Figure of the Earth,” Bull. National Research 
Council, No. 78 (February, 1931), pp. 186-87. 
396 
