6i6 



STUDIES IN FOSSIL BOTANY 



value of the characters which they have in common is 

 difficult to estimate. In some respects the Psilotaceae 

 seem to connect the Sphenop'hylls with the Lycopods, 

 but it is not wise to rely much on evidence from a 

 recent family in questions of remote ancestry. Although 

 various attempts have been made to group all the classes 

 of vascular plants under two main heads, it seems that 

 in the present state of our knowledge a threefold 

 division is still to be preferred, so as to avoid preju- 

 dicing questions which are still open. 



The affinities of the Equisetales, Pseudoborniales, 

 and Sphenophyllales are generally admitted ; to them 

 we may append the Psilotales, for reasons which 

 will be subsequently explained (p. 626). The Lyco- 

 podiales are at present best left by themselves, for the 

 fossil members show little or no approximation to any 

 other phylum. 



The Ferns likewise appear to have followed their 

 own course of development from the earliest times to 

 which our records extend ; they show, however, evident 

 affinities with the great bulk, at any rate, of the Seed- 

 plants. Our provisional arrangement may, therefore, 

 take the followine form : — 



Sphenopsida 



I 



Equisetales 



Pseudoborniales J- Articulatae. 



Sphenophyllales j 



Psilotales 



Lycopsida 



Lycopodiales 



Pteropswa 



Filicales 



Pteridospermeae 

 Gymnospermeae 

 Angiospermeae 



Spermophyta. 



