26 KENNETH HARTLEY 
The instrument must be very accurately leveled, that is, the axis of the 
instrument must be in the direction of the gravitational force. An angle of 
ninetv seconds between the axis of the instrument and the true vertical would 

Fig. 2. The gravimeter controls and adjustments. 
theoretically cause an error of one part in ten million, but it would cause a 
much greater difference in the actual measurement because of lateral strains 
in the filaments carrying the beam and mirrors. The present instrument is 

Fig. 3. The gravimeter in the field. 
easily adjusted to thirty seconds and that seems to be sufficiently close for 
the purpose. 
The dial readings give a direct measure of the differences in gravity and 
do not require any correction except to multiply by the calibration factor for 
170 
