Devonian Shales of Northern Ohio. 189 



lation. This conception is rather closely akin to, if not a direct 

 lineal descendant, of, Werner's old idea of "universal forma- 

 tions." Ulrich has himself called attention to the fact that this 

 view is widely at variance with the principles of sedimentation 

 laid down in such standard works as Geikie's, Kayser's, and 

 Dana's text-books of geology, and has taken exception to 

 Dana's rather axiomatic statement that "many a sandstone in 

 New York and Pennsylvania is of contemporaneous origin with 

 a limestone in the Ohio and Mississippi valleys." It may be 

 added that Mr. Ulrich's views on this subject are equally at 

 variance with the opinions of nearly all stratigraphers. Most 

 stratigraphers agree that the lateral variations of sedimentary 

 deposits are neither slight nor negligible. Although nowhere 

 explicitly stated, it is the reverse of this proposition which 

 forms the major premise in Mr. Ulrich's discussion of the Ohio 

 group stratigraphy. If, as this assumption indicates, the sedi- 

 ments of any Paleozoic sea were essentially the same from 

 center to shore, then a gray sandy shale such as the Chagrin of 

 northeastern Ohio could not possibly be synchronous in time 

 with the finer-textured black shales in the adjacent parts of 

 Ohio to the westward. This major premise, if admitted, would 

 of course require the assumption of axes and distinct ocean 

 basins to account for the juxtaposition of apparently synchro- 

 nous deposits of diverse lithologic type such as we find in the 

 paper under discussion. Concerning the soundness of this 

 fundamental assumption regarding the " persistence of litho- 

 logic units," it must suffice here to refer the reader to the 

 authorities cited above and to his own knowledge of the very 

 wide diversity of sedimentation now in progress in nearly every 

 lake and sea, regardless of the relief of adjacent lands. 



The maintenance of the doctrine of the " persistence of 

 lithologic units" has led Mr. Ulrich to make a far more exten- 

 sive application of the overlap hypothesis in treating problems 

 of correlation than any other author has permitted himself to 

 make. The overlap theory is of course in some cases a demon- 

 strahle fact ; in others, where only one or two or a very few 

 of the important facts are known, it often appears at first 

 to be a satisfactory explanation, but with access of knowledge 

 it i(f*many cases becomes inadequate. It is very easy to pass 

 a straight line through any two points but it may be impossible 

 to make it touch a third, and so with Mr. Ulrich's application of 

 the overlap hypothesis. When he attempts to make it cover 

 only two facts in the case, namely, deposits showing lithologic 

 and i'aunal differences and having a different areal distribution, 

 no difficulty is found in adjusting it to any such case where 

 other data are wanting, but when the attempt is made to fit it 

 to various other facts which must also be accounted for it may 



