266 THE GEOLOGICAL HISTORY OF PLANTS. 



Of the aboTe forms, fifty-one in all, found in the Brian 

 of eastern America, all, except the last four, are certainly 

 distinct specific types. Of these only four reappear in the 

 Carboniferous under identical species, but no less than 

 twenty-six reappear under representatiye or allied forms, 

 some at least of which a derirationist might claim as 

 modified descendants. On the other hand, nearly one 

 half of the Devonian types are unknown in the Carbon- 

 iferous, while there remain a very large number of Car- 

 boniferous types not accounted for by anything known in 

 the Devonian. Further, a very poor flora, including only 

 two or three types, is the predecessor of the Erian flora in 

 the Upper Silurian, and the flora again becomes poor in 

 the Upper Devonian and Lower Carboniferous. Every 

 new species discovered must more or less modify the above 

 statements, and the whole Erian flora of America, as well 

 as the Carboniferous, requires a thorough comparison with 

 that of Europe before general conclusions can be safely 

 drawn. In the mean time I may indicate the direction in 

 which the facts seem to point by the following general 

 statements : 



1. Some of the forms reckoned as specific in the De- 

 vonian and Carboniferous may be really derivative races. 

 There are indications that such races may have originated 

 in one or more of the following ways : (1) By a natural 

 tendency in synthetic types to become specialised in the 

 direction of one or other of their constituent elements. 

 In this way such plants as Arthrostigma and Psilophyton 

 may have assumed new varietal forms. (2) By embry- 

 onic retardation or acceleration,* whereby certain species 

 may have had their maturity advanced or postponed, thus 

 giving them various grades of perfection in reproduction 

 and complexity of structure. The fact that so many 

 Erian and Carboniferous plants seem to be on the con- 



* In the manner illustrated by Hyatt and Cope. 



