VELOCITY IN GRANITE AND LIMESTONE 109 
METHODS 
The section of granite selected for this work had a reasonably level 
and continuous surface for slightly more than 400 feet. No over- 
burden of any sort was present. 
A recording set-up comprising five detectors spaced 25 feet apart 
was placed at one end of this section of granite. It was found that 
it was. quite essential to fasten the detectors rigidly to the granite 
since, when this was not done, relatively large errors were introduced 
by poor coupling. In order to accomplish this, plaster-of-Paris dams 
were built around the detectors and melted sulphur poured in. After 
the sulphur had solidified, the detectors were found to be cemented 
rigidly to the granite. Since it was necessary to use such extreme care 
in the mounting of the detectors, they were not moved during the 
course of the profile, but the shot point was moved instead. Shot 
points were placed every too feet, from 5 feet to 305 feet from the 
nearest detector. This gave data at 25-foot intervals, from 5 to 405 
feet, with a check or overlap position on each record. 
No indication of the instant of detonation was used in this work 
because of the possibility of small errors being introduced from this 
source. Times were measured from the instant of arrival of the wave 
to the nearest detector. 
Two complete sets of data were taken on this profile. The first 
set was taken with a normal film speed in the oscillograph, that is 
with o.or second covering a distance on the record of about 0.75 
centimeter. The second set was taken with a high film speed in the 
oscillograph, making the same time interval cover approximately 2.50 
centimeters. 
This procedure was carried out for both the longitudinal waves 
and the transverse waves. Separate sets of records were obtained for 
each type of wave. Specimens of the high and low speed longitudinal 
and transverse wave records are shown in Figure 2. 
RESULTS 
From an examination of Figure 3, it is seen that the time-distance 
graphs for both the longitudinal and transverse waves are slightly, but 
unmistakably, curved. If the observed points are plotted on a scale, 
of one half that shown in Figure 3, the graph then appears to be a 
straight line and it is only when an enlarged scale is used that the 
curvature becomes apparent.‘ From this it is evident that there is an 
‘ Proof of the curvature of this line may be had by a least-square solution. Assum- 
. ing a straight line, the residual is 0.00138 second while with the curve it becomes 
0.00013 second. 
631 
