METHODS OP CORRELATION 465 



seems to me that the evidence derived from these two or three sources 

 should not have the same weight. It is not that the evidence of either is 

 distinctly inferior to that of the other, but that the two kinds of data — ■ 

 land and marine — are necessarily more or less discordant in their re- 

 spective tendencies. The one points to or centers in emergent stages of 

 geologic history, the other in submergent stages. Consequently, in view 

 of the fact that the time classification of geologic events is based pri- 

 marily on displacements of the strand-line — in other words, on criteria 

 furnished by marine factors — it follows that there is no such thing as a 

 distinctly Devonian land flora having a significance corresponding, in the 

 matter of time and duration of existence, to that conveyed by the words 

 "the Devonian marine fauna." The latter is most typically expressed by 

 its median, Onondaga-Hamilton f acies ; hence by a f aunal concept mark- 

 ing a time when the land flora would naturally be at its minimum for 

 the period and probably in course of adaptation to changing conditions — 

 in transition from the facies that prevailed before to that which flourished 

 later. The successive facies of the land flora thus must have attained 

 their respective most typical developments in the intervals between those 

 times when the successive marine faunal facies reached their respective 

 high points. The two may be said to have alternated. 2 



BEARING- ON THE AGE OF THE OHIO AND CHATTANOOGA SHALES 



Considerations like these may largely explain the instability of the 

 Devonian-Mississippian boundary in geologic literature. In areas con- 

 taining beds rich in land plants and poor in marine fossils, the decision 

 of the paleobotanist, providing his judgment is determined solely by pre- 

 ponderance of the floral affinities, would be likely to favor alliance of 

 really early Mississippian deposits with the underlying true Devonian. 

 The invertebrate paleontologist, on the contrary, especially in areas where 

 marine fossils are plentiful, would be impressed by certain changes in his 

 faunas, and therefore more likely to favor placing the doubtful beds into 

 the younger system. 



2 The principle here brought out was submitted for critical comment to Doctors F. H. 

 Knowlton and Arthur Hollick, the two most experienced and doubtless best qualified of 

 our authorities on Mesozoic and Cenozoic floras. It was recognized that the test would 

 be severe and the result perhaps indecisive, if not distinctly unfavorable, because the 

 post-Paleozoic ages were characterized much more commonly by emergent conditions, 

 hence by wide prevalence of land, than were the Paleozoic ages in which the submergent 

 phases are generally thought to have predominated. On the contrary, however, I am 

 glad to be permitted to report the result of their investigations as distinctly favoring 

 the hypothesis. As a rule, they found that, beginning with the Permian, and thence, on 

 to the present time, the terminal floras of each period are more readily distinguishable 

 than is either one from the nearest flora of the preceding or the succeeding period, as 

 the case may be. For instance, the early Jurassic flora is much more easily distinguish- 

 able from the late Jurassic facies than it is from the preceding late Triassic flora. The 

 late Jurassic flora, again, is considerably like the early Cretaceous flora. 



