108 WEATHERBY, BORN, AND HARDING 
waves or Rayleigh waves, were observed on these records, they could 
not be interpreted with sufficient precision to draw any definite con- 
clusions from them. Since it is necessary to obtain the transverse 
velocity in order to determine the elastic constants of the rocks by 
the seismic method, this profile shed no light on the problem. 
The second profile, extending over a span of 400 feet, was laid 
out on a section of the granite which was free from fissures. The rec- 
ords taken over this profile showed both the longitudinal and trans- 
verse waves. 
INSTRUMENTS 
It was evident that it would be necessary to measure relatively 
small increments of time with a high degree of accuracy if a large 
number of points were to be obtained in this interval of 400 feet. If it 
be assumed, for the moment, that the velocity to be measured is of 
the order of 16,000 feet per second, and that the distance over which 
a velocity is to be determined be set at 100 feet, it would then be 
necessary to measure a time interval of the order of 0.0063 second. 
Since an accuracy of 2 per cent was desired, it was necessary to design 
recording and measuring instruments capable of detecting time dif- 
ferences as small as 0.00012 second. 
The desired precision was attained by exercising extreme care in 
the selectton of a perfectly matched set of detectors and recording 
channels.* The optical system was designed to give maximum sensi- 
tivity with a minimum width of the beam of light reflected from the 
galvanometer, thus producing a fine, clean-cut line on the record. The 
high frequency response of the recording systems was made as great 
as possible and a special fine-grain recording paper was used. The 
result was an extremely sharp, clean-cut, high amplitude break in 
the line corresponding with the time of arrival of the longitudinal 
energy. Time intervals of 0.01 second were placed on the record by 
means of an electrically driven tuning fork and a synchronized vibrat- 
ing reed. This interval did not vary from an absolute value of 0.01 
second by more than 0.04 of 1 per cent. Two complete sets of data 
were taken on this profile, one with a high speed oscillograph and the 
other using normal film speed. The purpose of the high film speed was 
primarily to obtain additional accuracy. The records were read with 
the aid of a micrometer microscope with which it was possible to read 
distances on the records corresponding to 0.00002 second. 
3 Theoretically this is not necessary where impulses corresponding with the first 
arrival of energy are observed. This added precaution was taken to insure uniformity 
in the later portions of the records. 
630 
