PROPER USE OF FOSSILS IN CORRELATION 489 



and ostracods, or, better still, nodes in some cases and pits in others on 

 tlie middle of the head and on free cheeks; or spines on the head, tail, 

 and pleura which occur in endless variety of number, length, and ar- 

 rangement ; in the case of the conodonts and annelid jaws, which often 

 are found in beds otherwise nearly barren, excellent criteria of this kind 

 are offered by peculiarities in the number, arrangement, length, and thick- 

 ness of the denticles. Though any of these features may be repeated 

 more or less precisely in other species of the same genus as well as in 

 other genera of the same class, combinations of two or more minor pecu- 

 liarities are commonly notable by which the several occurrences may be 

 distinguished, and these combinations are never exactly reproduced. 



Extravagance in development of any feature could not have lasted long ; 

 hence species so distinguished are reliable horizon markers. Even the 

 Ostracoda, referring particularly to the more or less complexly marked 

 Beyrichidse and Cytheridge, are of great value in identifying horizons. 

 Valves and entire carapaces of these small Crustacea often are exceedingly 

 abundant, widely ranging, and commonly better preserved than are the 

 remains of other classes of fossils. I found them most useful and reliable 

 in correlating the Silurian and early to Middle Devonian formations in 

 the Appalachian region. 



Composite types like the Bryozoa, which have many individual and 

 colonial characters,- are especially useful, because biologically unimportant 

 differentiations may yet, and commonly do, produce positively recogni- 

 zable combinations. The Bryozoa have been of the greatest help in corre- 

 lating Ordovician and Silurian horizons; and now the detailed studies of 

 Canu and Bassler are proving them to be equally valuable in establishing 

 the time relations of Tertiary deposits in the southern Coastal Plain of 

 the United States. 



By the recognition of such minute structural differentiations we actu- 

 ally identify geological horizons, for it seems practically impossible that 

 the same combination of minor and major biological characters can have 

 existed either very long or more than once. In my opinion, therefore, 

 the occurrence of a single finely drawn variety in two or more widely 

 separated places— be they in the same province or not— is a more trust- 

 worthy indication of the contemporaneity of the beds containing it than 

 would be any quantity of the indefinite testimony afforded by generic 

 alliances. 



Summing up, the relative finality of our paleontological work in cor- 

 relation depends (1) on how closely we discriminate our fossils, (2) on a 

 proper appreciation of the relations of the evidence of marine animals to 



