STRATIGRAPHIC PALEONTOLOGY. 93 



Schimper, the Oligocene containing 10 to 15 per cent of living species of 

 Mollusca and the Paleocene practically none. The comparison of faunas 

 by means of such percentages is likely to lead to very serious errors and 

 possesses significance only when all the species of a given ecologic complex 

 are known and are compared with all the species of a similar ecologic com- 

 plex. In the discussion of fossil faunas too little attention has been paid 

 to differences in contemporaneous faunas due to differences in environment. 

 The necessity of giving greater consideration to ecology in the study and 

 geologic use of fossils has been recently emphasized by Vaughan. 1 If the 

 percentage method of comparing faunas were applied to the living lagoon 

 fauna and the living exposed barrier reef fauna of Cocos-Keeling Islands, 

 on the assumption that the lagoon fauna is geologically Recent, the barrier 

 fauna would be Miocene, and the fauna of the pools and flats behind the 

 barrier reef would be oldest Pliocene. By this method much of the West 

 Indian Miocene might be made Pliocene or even Pleistocene. Errors may 

 be introduced because a fossil fauna may not be homogeneous. Marine 

 Mollusca, for example, may be collected from different ecologic stations 

 at a single locality within a single formation. Mud-burrowing and boring 

 bivalves have been washed upshore and mixed with gastropods that lived 

 only on intertidal sand flats ; offshore and estuarine forms have been washed 

 in among indigenous shoal-water species. Yet percentage comparisons 

 are valuable if they are applied to complete faunas of a similar ecologic 

 complex, and though one may seldom be sure of having collected all the 

 members of any fauna, careful and thorough collecting may give a large 

 proportion of it. The biologic method of correlation affords its best results 

 only if it is applied after a study of the vertical distribution of species, 

 particularly the minor variants of species, in accurately ascertained colum- 

 nar sections. By studying in this way the faunas and the associated sedi- 

 ments over areas of moderate extent the stratigraphic range of many species 

 and variants may be ascertained, and these species rather than percentages 

 supply a basis for the correlation of similar deposits. Even this method of 

 work involves liability to error, however, because the observed strati- 

 graphic ranges of the organisms studied may not sufficiently represent the 

 actual ranges. 



In establishing the correlations of the formations of the Dominican 

 Republic all three methods were used. It was not practicable in many 

 places to trace the formations from one area to another even where it is 

 possible to do so; the record of earth movements, although it has not been 

 fully deciphered, was a valuable aid, but the most generally useful method, 

 and that by which the ages of most of the formations were inferred, was 



1 Vaughan, T. W., Fossil corals from Central America, Cuba, and Porto Rico, with an account of the 

 American Tertiary, Pleistocene, and Recent coral reefs: U. S. Nat. Mus. Bull. 103, pp. 190-193, 1918, 

 (Section entitled "Geologic correlation by means of fossil corals.") Corals and the formation of coral reefs: 

 Smithsonian Rept. for 1917, pp. 186-276, pis. 37, text-figs. 16, 1919. 



