REFLEXION OF ELASTIC WAVES 145 
event on one seismogram was correlated with the shallower event on 
the other; such errors may be of any magnitude. Errors of this nature 
are usually detected by the unreasonable structural interpretation 
resulting. It is, however, possible to correlate bursts of energy cor- 
rectly between two seismograms and yet to choose different points as 
representing the commencement of the event in the two cases, result- 
ing in the correlation being most probably in error by one complete 
cycle or wave length which may represent an error in computed depth 
of from 50 to 150 feet, depending on the average velocity of propaga- 
tion and other factors. This is a common error and can best be elimi- 
nated by striving to obtain good quality seismograms in which the 
beginning of a burst of energy is readily identified. Where several 
horizons are being followed simultaneously and local geology leads 
one to believe their interval will remain uniform, this error is obvi- 
ously under much better control, since it would disclose itself as ap- 
parent changing of interval between horizons being followed. 
Errors in time measurement are entirely a matter of instrumental 
design and can readily be reduced to 0.1%, which rarely represents an 
error in computed depth of more than 10 feet. The most serious factor 
is usually the method used in recording the instant of explosion of 
the dynamite charge, as electric blasting caps frequently exhibit a dis- 
concerting irregularity on time of explosion under differing con- 
ditions. 
Errors in computation of depths mainly arise from incorrect as- 
sumed velocities of propagation for the various strata traversed by 
the wave front. Figure 6 represents a typical stratigraphic column in 
Oklahoma and also shows the velocity of propagation of the longi- 
tudinal wave in the various members. Tracing the velocity zones 
upwards from the Viola limestone we first encounter a series of lime- 
stones interspersed with shales, which zone has a rather high average 
velocity of propagation, being approximately 15,000 feet per second. 
The thick overlying series of Pennsylvanian shales and thin limestone 
beds is found to exhibit a velocity which decreases with decreasing 
depth, the range being about 13,000 to 9,000 feet per second. A zone 
some 100 feet thick which may possibly be caused by the ground 
water table is found to have a rather lower velocity of about 7,000 
feet per second, and finally we have the surface zone, which is, I 
believe erroneously, sometimes referred to by seismologists as the 
“weathered” surface layer. This superficial zone in Oklahoma has a 
velocity of propagation of some 2,000 feet per second. My objection 
to the description ‘‘weathered” layer is that where the surface is cut 
by a ditch adjacent to a shot point I have many times observed that 
805 
