100 THE AMERICAN BOTANIST 



It seems right therefore to conclude that a true classifi- 

 cation must hnd its chief justification in its harmony with fos- 

 sil evidence, as well as with all other kinds of evidence. One 

 may even assert that no evidence is so unquestionably con- 

 clusive as is that drawn from fossils. 



In applying these considerations to the old established 

 group Pteridophyta, we find a curious jumbling of forms with 

 superficial resemblances. There is ample reason to believe 

 that' the phylogeny of these groups runs somewhat along the 

 lines of our figure. 



There is of course some logic still in uniting all of the 

 seedless V'asculares in one group, Pteridc^phyta, and the seed 

 bearing forms in a separate and co-ordinate group Sperma- 

 tophyta. But can these groups properly be called co-ordinate? 

 Is not this to becloud the facts of anatomy and palaeontology? 



No one familiar with fossil botany ([uestions the intimate 

 relation of ferns (Aspermae) and gymnosperms. We have 

 had proposed the intermediate groups Cycadofilices and Pteri- 

 dosperms. There is on the contrary no known connection be- 

 tween the Lycopsidan and Pteropsidan branches of Tracheate 

 plants. This fact is just as cogent if we accept Scott's 

 Sphenopsitla as co-ordinate with Lycopsida and Pteropsida. 

 The logical separation therefore occurs at this point of 

 "unconformity." The basic division of Vasculares is into 

 Lycopsidan and Pteropsidan (and perhaps Sphenopsidan) 

 branches. In this case Pteridophyta and Spermatophyta no 

 longer hold. D. H. Scott writes : "It will be noticed that this 

 grouping crosses the customary division of Vasculares into 

 Pteridophyta and Spermatophyta. Though the traditional 

 classification will, no doubt, continue to be used on grounds 

 of convenience, it no longer, in the light of the palaeontologic- 

 al evidence, expresses a natural arrangement, tor the affinities 



