INTERNATIONAL GEOLOGIC INSTITUTE 605 



the carbon dioxide of the atmosphere has been mainly sup- 

 plied concurrently with its consumption, and the amount has 

 varied with the ratios of supply to consumption, and these ratios 

 have been dependent upon the extent of the land and the condition 

 of the sea, and if further the impoverishment of the'atmosphere in 

 carbon dioxide is determinable by the unusual prevalence of cer- 

 tain kinds of deposits, as salt, gypsum, red elastics, and glaciated 

 bowlder clays, while its enrichment is indicated by equable tem- 

 peratures and mild climates in high latitudes, there is ground for 

 hoping that the constitution of the atmosphere may be made to 

 afford a valuable means of correlation applicable when the 

 paleontologic criteria are most liable to fail. Effects due to the 

 constitution of the atmosphere must be universal and strictly 

 simultaneous, though of course not everywhere identical. For 

 example, in India, Australia, and South Africa, an abrupt change 

 in the flora took place at some time near the transition from the 

 Carboniferous to the Permian period. Fontaine and White have 

 found an abrupt, though less radical, change in the flora of east- 

 ern America at about the same time, but there are no paleonto- 

 logical means at present known by which these changes can be 

 accurately correlated. The change in India, Australia, and 

 South Africa is closely associated with glacial deposits. Now, 

 if glaciation be a result of a constitutional state of the atmos- 

 phere, such state should make itself felt in all parts of the earth 

 simtdtaneously, though in different ways and degrees, and the 

 floral change in eastern America could with good reason be 

 strictly correlated with the changes in Asia and the southern 

 hemisphere, and should find verification in similar changes else- 

 where. This is merely a hypothetical illustration. The sup- 

 posed atmospheric mode of correlation should be verifiable by 

 its peculiar effects, for it should simultaneously affect distant 

 floras made up of different constituents, a phenomenon differing 

 in nature from the effects of simple migratory replacement or 

 biologic evolution. 



It is obvious, however, that correlation by atmospheric 

 states can only become a trustworthy dependence by the 



