REFLEXION OF ELASTIC WAVES 141 
controlled to reduce the probable error in relative depth measure- 
ment over short distances to between ro and 25 feet even when con- 
touring a horizon at 5,000 or 6,000 feet in depth. Our first interest, 
therefore, is in the quality of seismogram which can be obtained and 
the certainty of correlations made between adjacent points. Figure 5 
(a) to (d) are reproductions of typical seismograms from various 
areas. 
Figure 5 (a) is a seismogram made by the McCollum Exploration 
Company in central East Texas. The prominent reflexion occurring 
at about 0.93 second is from the Austin chalk, a limestone member of 
the Upper Cretaceous which can be identified on reflexion seismograms 
throughout this area. 
Figure 5 (6) shows a Nacatoch reflexion obtained by a Geophysical 
Service Inc. unit in the extreme north-east corner of the East Texas 
basin. The Nacatoch is a bed in the Navarro formation of Upper Cre- 
taceous age consisting of sands and shales witha limestone member 
which is no doubt responsible for the reflexion. 
Figure 5 (c) is a typical Oklahoma seismogram recorded by a 
Geophysical Service Inc. unit in central Oklahoma. The Oswego and 
Mississippi limestone reflexions occur at 0.72 and 0.87 second respec- 
tively. These beds, or their equivalents can be followed throughout 
central Oklahoma and parts of Kansas; they are responsible for the 
outstanding success of this method in the Mid-Continent area. 
Figure 5 (d) reproduces a series of seismograms made by a Seismo- 
graph Service Corp. unit in northern Oklahoma. The event occurring 
on all these records at about 1.20 seconds is the Mississippi limestone 
reflexion on which correlation was based. This series of records demon- 
strates the degree of certainty normally achieved in correlation of 
seismic events in this area. 
From a study of these seismograms I think it must be admitted 
that under a variety of conditions in widely scattered areas we are 
able to comply with the mechanical requirements of the method. We 
must now question the accuracy of depth determinations made from 
these seismograms. 
ERRORS IN DERIVED DATA 
Sources of error fall in three categories: (a) errors in correlation, 
(5) errors in time measurement, and (c) errors in computation of 
depths. 
Errors in correlation may be of two kinds: the first would arise if 
two bursts of energy were present in an area as in the case of the seis- 
mogram reproduced in Figure 3, and at adjacent points the deeper 
801 
