SHORTER CONTRIBUTIONS TO GENERAL GEOLOGY, 192L 



LITHOLOGIC SUBSURFACE CORRELATION IN THE "BEND SERIES" OF 



NORTH-CENTRAL TEXAS. 



By Marcus I. Goldman. 



OBJECT AND METHODS OF THE INVESTIGATION. 



This paper presents the results of an attempt 

 to obtain, by a study of well drillings, criteria 

 for correlation, by the character of the beds 

 encountered, of horizons in the "Bend series" 

 of north-central Texas. The result of the 

 work most directly applicable to the needs of 

 the oil-well operator is that of enabling him 

 to know more exactly the ])osition of the 

 producing beds in his well — a result that is of 

 particular value in north-central Texas because 

 certain beds in that region which give no show 

 of oil to the drill may on being "shot" yield 

 considerable oil. To know the part of the 

 stratigraphic section reached by the ch'ill at 

 any stage of the boring also enables the driller 

 to be on the alert for indications of oil at 

 horizons where productive sands are encoun- 

 tered in other wells anil to know when the drill 

 has passed all horizons at which the beds are 

 likely to yield oil. 



In two wells that lie near each other the 

 succession of beds is so nearly the same that 

 at least some horizons can be correlated from 

 the driller's logs if they are carefully kept. 

 But the character of any bed may gradually 

 change from place to place, so that the farther 

 apart any two wells are the more difficult it 

 becomes to recognize the beds that lie at the 

 same geologic horizon in them. In undertak- 

 ing this study, however, I started with the 

 assumption that, though (he composition of a 

 bed at a certain horizon in any two widely 

 separated localities might be so dilferent that 

 the bed could not be recognized as such, the 

 relation of its composition to the composition 

 of the beds above and below it would show 

 enough similarity to jjcrmit its identification. 



That is to say, it seemed probable that any 

 chano:e in the character of the sediment deliv- 

 ered to one part of an open, continuous basin 

 of deposition would be manifestetl in all other 

 parts of th(> basin. Thus, an increase of sand 

 would be recorded all over the basin, in one 

 locality perhaps by an increase in the sandiness 

 of a limestone, in another by the deposition 

 of coarse sandstone in the mid^t of shale, and 

 in still another merely by a slight increase in 

 the amount or the size of sand in a shale, but 

 everjrwhere there would be an increase in the 

 proportion of sand; and the same rule would 

 apply to the other constituents.' To bring 

 this out it was necessary to represent for ea(^h 

 sample obtairnMl from a well the jiroportions 

 of the three ])rincipal ingredients — sand, clay, 

 and lime. Lime as here used includes all the 

 transparent carbonates of the calcite group. 

 No attempt to differentiate the several mem- 

 bers of that group was made, though the 

 dilferent iation of calcite and dolomite, for 

 example, would be valuable. The lower limit 

 for size of grain of material classified as sand is 

 about 0.0.5 millimeter, but this limit is not 

 delinite, depending somewhat on factors other 

 tlian size. If material as fine as 0.02 millimeter 

 were abundant in a thin section it would be 

 estimated as sand, but beds containing material 

 of that fineness are likely to appear essentially 

 as clay. 



The most convenient metliod of compiling 

 the results obtained seemed to be what I have 

 called the "percentage log," shown in the 

 third and fourth columns on Plate I (in pocket ). 



■ Though apphcd much more roughly, this principle is similar to 

 (hat used by De Geer in correlating the glacial clays of Sweden. See 

 De (ieer, Gerard, A gcochronology of the last 12,000 years: Cong, gi^ol. 

 intcnuit., U" sess., Compt. rend., fasc. 1, p. 241, 1910. 



1 



