INTRODUCTION 455 



but the knowledge that the shallow-water faunas of the present seas are 

 very different from those living on the bottom at depths of 1,000 feet or 

 more is of little practical value to the stratigraphical paleontologist. It 

 is of little value to him, because the fossiliferous beds which he is called 

 on to study or correlate have been, so far as I know, without a single well 

 established exception, laid down in waters that varied too little in depth 

 to have suffered much change in faunas on that account. 



PRECEDING PAPERS ON THE SUBJECT OF CORRELATION 



On at least two occasions my views on the criteria and principles of 

 stratigraphic correlation have been expressed in considerable fullness. 

 The first of these appeared in my Eevision of the Paleozoic Systems, pub- 

 lished by the Geological Society of America in 1911 ; the second in a paper 

 delivered before the International Geological Congress at Toronto in 1913, 

 and since published in the proceedings of the session. The former, a 

 work of 400 pages, is devoted almost entirely to the discussion of the 

 criteria and principles, those pertaining to the lithologic and other phys- 

 ical aspects of the problem being treated quite as fully as the paleonto- 

 logic. The second paper, covering 75 pages, begins with a concise state- 

 ment of the principles of correlation, with special reference to strati- 

 graphic taxonomy. The object chiefly in mind on this occasion was to 

 illustrate the practical and scientific value of the diastrophic methods of 

 classification, as defined and recommended in the larger work, by its appli- 

 cation in a well known and supposedly difficult case, namely, the Ordo- 

 vician-Silurian boundary in America. 



The essentials of the present effort are contained in these published 

 papers. But, strictly speaking, there is little of repetition in these papers, 

 and what there is pertains wholly to the statement and general discussion 

 of principles and criteria. The illustrations are new and some have de- 

 veloped only daring the past year or two; in fact, there is no lack of 

 examples that may be cited as proving the inadequacy and imperfection 

 of correlation methods commonly practiced. Their abundance, more- 

 over, forces the regrettable conclusion that the errors which have resulted 

 from the prevailing practice are more the rule than the rare exceptions 

 which they should be. 



NEW CRITERION IN CORRELATION 



Although the greater part of the following discussion may be described 

 as destructive criticism, I trust the reader will find also much that de- 

 serves to be characterized as constructive. Comprised in the latter is a 

 new organic criterion that is of exceeding and obvious value. Its value 



