PRINCIPLES OF STRATIGRAPHIC CORRELATIONS 523 



they have been deposited within the same continental basin or in more 

 exceptional instances when they fall within the broader limits of a faunal 

 province. Beyond such limits exact contemporaneity can not be estab- 

 lished absolutely, however similar the beds may be in petrological char- 

 acters. Approximate contemporaneity in such cases is assured only 

 when such beds are bounded above and beneath by more widely trans- 

 gressing stratigraphic units or faunal zones of not very different ages. 

 In the absence of such determining zones essential contemporaneity may 

 sometimes be established by diastrophic criteria. 



Using a large unit the principle is illustrated by the lithologically 

 varying representatives of the Waverlyan system in the three or more 

 subdivisions of the Ohioan province recognizable in this period. Thus, 

 the equivalents of the sandy and shaly formations of Waverlyan age in 

 the Appalachian Valley north of Tennessee and east of the Cincinnati 

 axis can not be determined by lithological criteria in the cherty Tulla- 

 homa and Fort Payne limestones of Tennessee and Alabama, nor in the 

 limestone and calcareous shale deposits of this period in the Mississippi 

 Valley above Saint Louis. Within each of these subprovinces the lithic 

 features of the formations remain fairly constant, so that correlations 

 by similarity of such characters is practicable over areas of considerable 

 extent. But beyond these correlations are possible only by means of 

 diastrophic criteria indicated by changes in sediments and faunas. Using 

 smaller units we have good illustrations (1) in the Kimmswick lime- 

 stone, which has been erroneously correlated with the Galena dolomite 

 because both are preceded by limestones of Lowville age (^%wer Tren- 

 ton") and succeeded by Maquoketa shale; (2) in the Lowville limestone 

 in east central Tennessee, which was classified as Carter limestone be- 

 cause both are underlain by the Lebanon limestone and overlain by the 

 Hermitage shale; (3) in the Izard limestone of Arkansas, erroneously 

 correlated by Ulrich with the Viola limestone of Oklahoma, and (4) in 

 the Bertie waterlime, which was referred to the Eondout until it was 

 proved an older bed. 



(4) Correlation of dissimilar lithologic units. — Difference in lithologic 

 characters of beds that are not continuously exposed, but occupy similar 

 positions in the stratigraphic column, is always suggestive of difference 

 in age, providing the faunas are not the same. If both the fauna and 

 the kind of rock remain unchanged in a given broad region, and if both 

 factors are appreciably and uniformly different in another wide area, 

 one of two conclusions is to be drawn : either the two kinds of rock are 

 of the same age and the differences are due solely to provincial distinc- 



