the corresponding sand sections but also a close resemblance between the pat- 

 terns of the two curves. Note, for example, such minor features as the slightly 

 sandy shale streaks at 3666 feet and 3691 feet, which are represented as slight 

 bulges in the potential curve. The characteristics of electric and drilling-time 

 curves are determined by changes in lithology. The drilling-time break at 3676 

 feet is interpreted as anomalous to lithology because a connection was made at 

 this depth and part of this foot was drilled with the clutch out, causing an 

 erroneously timed foot to be registered. By marking on the log where interrup- 

 tions in drilling occur, such features may be recognized with ease and incorrect 

 interpretations prevented. 



The curves of Figure 18-3 which were plotted from the same data as those 

 of Figure 18-2, show by solid lines the loss of detail in using five-foot intervals 

 instead of one-foot intervals and by dotted lines the effect of averaging when 

 using drilling-rate values. In the upper solid curve on the correlation scale the 

 total time for five feet was used and plotted as a bar curve according to the 

 manner generally practiced on sample logs. In the lower solid curve on the 

 detailed scale, the average time per foot was plotted as a point-to-point curve. 

 In both of these curves the presence of two sands and one shaly sand is observed, 

 but the exact depths at which they occur, their net thickness, and minor litho- 

 logic breaks are absent. Obviously, it requires more time to plot the curves in 

 Figure 18-2 than in Figure 18-3, and the information to be gained is dispropor- 

 tionate to the time saved. There is some value in large-interval drilling time 

 and in drilling-rate curves to be sure, but their use is restricted to problems 

 where only general impressions are needed either for correlation or lithologic 

 interpretation. In plotting sample logs on the basis of percentage of lithologic 

 types present in each sample, the position of major breaks may be determined 

 by plotting a drilling-rate curve similar to the upper dotted curve in Figure 18-3. 

 Where difficult full-length correlation problems are encountered, however, the 

 drilling-time curves of Figure 18-2 will be found far more reliable and useful. 

 Drilling-rate logs are useful in sample examination work in determining sample 

 lag, but here again the information is only exact within the limits of accuracy 

 of the method used. Further discussion of the application of drilling time and 

 drilling rate follows at the end of the next section. 



Several devices purport to record changes in rate of penetration in terms 

 of feet per hour, and mechanical instruments have been marketed that provide 

 drilling-time logs or data from which these logs may be plotted manually. One 

 of the drawbacks to the use of drilling-time logs has been the time required to 

 plot curves of the type illustrated in Figure 18-2. There is no machine available 

 that will reliably record or plot drilling-time logs of this type and eliminate hu- 

 man errors. Considerable experimental work along this line has been done, and 

 the need for such a device is great. 



373 



