388 H. S. WILLIAMS CORRELATION PROBLEMS 



method, which I presume paleontologists in general use in determining 

 correlation of formations. In my case I have adopted several rules which 

 I regard as of prime importance, namely : 



1. To establish correlation between two fossil faunas it is essential to 

 bring into comparison identical, not closely related species. 



2. It is essential in listing the species of a fauna to include only those 

 species which are actually found associated together in the same stratum 

 in the same locality, or when several local faunules are combined only 

 those local faunules may be listed together that have combined enough 

 of the same species to certify that they are samples of the same fauna. 



This precaution is required to eliminate such differences as may arise 

 by alternation of one or more faunas in the same general formation, or 

 by change in the original environmental condition often expressed by 

 differences in the lithological constitution of the sediments, or by differ- 

 ences arising from geographical distribution. 



Other rules may have been applied, but these two are essential if we 

 expect to reach precision in correlation. 



Having in this way built up a list of the contents of the Edmunds 

 fauna of Maine, in order to apply the same rules in the transatlantic 

 fauna compared, I found the typical Wenlock series of England is ex- 

 pressed in the Shropshire section, as reported in Davidson's Supplement 

 of British Fossil Brachiopods, which has been treated with the same kind 

 of precision so far as Brachiopods are concerned. This is taken as th« 

 standard of comparison for the English sections. In making comparison 

 with the general Niagara fauna of the country I have used the fauna oi 

 the Bochester shale of New York as the standard of comparison. 



In applying the statistical method I have adopted the following rules 



3. To establish correlation, species of the same class must be com- 

 pared. In the present case the class of Brachiopods are used because th( 

 statistics regarding that class have been most thoroughly elaborated. 



4. The particular horizon in the standard section with which equiva- 

 lence is found in the compared section is determined by the number of 

 identical species. 



Thus in the case before us, comparing the Edmunds fauna with the 

 general Silurian faunas of Great Britain, I found 18 species identical 

 with species of the transatlantic Silurian; 17 of them are listed in the 

 Wenlock limestone and shale, 3 do not appear above the Wenlock lime- 

 stone, 16 appear in the Woolhope, 14 and 1 doubtful in the Upper Llan- 

 dovery, 12 appear in the Ordovician, namely, in the Lower Llandovery 

 and Caradoc, and 4 as low as the Llandeilo. Above the Wenlock 12 ap- 



