392 H. S. WILLIAMS CORRELATION PROBLEMS 



The personal equation of the describer of the species listed must be 

 eliminated before we can reach exact conclusions. 



For instance, Barrande, who prepared a monumental set of volumes 

 illustrating the Silurian system of Bohemia, made a list of 49 species of 

 Spirifers in the one fauna E 2 . These volumes were published in 1879. 



In Murchison's Silurian System of Great Britain, published in 1839, 

 Sowerby described the fossils and recognized only 12 species under that 

 generic name — Spirifer. The Brachiopods were later thoroughly studied 

 by Davidson, and in 1883, after a life study of them, his final list of all 

 the British Silurian Brachiopods known to him when the final supple- 

 ment was published (in 1883) contains only 5 species (and 3 varieties of 

 1 of the species) of the genus Spirifer. 



Only one of the names used by Barrande was a species previously rec- 

 ognized by another author ; 48 of the 49 species he labeled as new species. 



In Davidson's list 3 of the species recognized were described by for- 

 eigners before Sowerby, 2 of them by Sowerby ; 1 is a species of McCoy, 

 and 1 of the varieties is by Salter, both of the latter working chiefly on 

 British species. 



Although the lists would seem to indicate that the Bohemian fauna 

 was extremely rich and unique in species, examination of the figures 

 shows that Barrande's 49 species of Spirifer, of E 2 Bohemia, belong al- 

 most entirely to modifications of the 5 species listed by Davidson from 

 the British Silurian. There are some peculiar forms not recorded bj 

 any of the figures of Davidson; but the Bohemian series presents vei 

 little morphological digression from the group of forms described as 

 species by Davidson. 



This is but an extreme case of the different conceptions of species held 

 by authors. 



Not only is there great difference in the plasticity of characters of 

 natural groups of specimens listed under a single specific name, but there 

 is a great difference among authors in the application of specific names 

 to the specimens under examination. 



This analysis shows that a list of identical species by which we attempt 

 to correlate the faunas of two geographically separate formations differ 

 widely in many respects, namely: (1) in the amount of morphologic 

 divergence compatible with specific identity; (2) in the range or strati- 

 graphic thickness of the sediments through which they are known to 

 persist; (3) in the distribution or wideness of the geographical area in 

 which they may be found, and, what is of still greater importance in 

 estimating their value for correlation, (4) in the wide divergence of 



>m 



s 



