THE QUEST OF PHYLETIC LINES 105 



difficult. All such assumptions must be submitted to the check 

 of physiological probability, while their risk as pure assumptions 

 may be diminished, or even removed, by the recognition of inter- 

 mediate steps. 



2. The neglect of collateral checks. In order to place a phyletic 

 view on a footing of reasonable probability it will not suffice to 

 advance a single train of facts. The fossil record was not taken 

 into account in checking the assumption that the Leptosporangiate 

 Ferns were the earliest, nor in assuming that the Lycopodium 

 stele was the result of the fusion of the numerous steles of Selagi- 

 nella. These conclusions fell away so soon as the palaeontological 

 check was applied. It is not uncommon, especially in the work 

 of beginners, to see broad phyletic conclusions drawn upon the 

 basis of merely anatomical facts, or of details of a sporangium, 

 or of the contents of an embryo-sac, without any reference to 

 the broader features or history of the organisms compared. All 

 such conclusions must be held as provisional, and open to revision 

 whenever a broader basis of comparison is adopted. 



3. Neglect of the fact that intermediate steps between the organisms 

 compared are not known. The prominent example of a compar- 

 ison wrecked on this ground is that of the Hymenophyllaceae and 

 the Mosses. No evidence of any intermediate step between these 

 divergent types could be produced. On the other hand a line of 

 probability emerges from the detailed study respectively of the 

 Musci and of the Filicales, that the characters which formed the 

 basis of the comparison are consequences of specialisation along 

 parallel lines. This mode of reasoning is the source of many of 

 the fanciful and even divergent opinions which have from time 

 to time been promulgated with regard to the morphology of the 

 distinct phyla of Pteridophytes. It is in the first instance 

 assumed that they must conform to a certain structural mould. 

 It is forgotten that intermediate steps do not now exist between 

 them, and the attempt is then made to force the divergent types 

 into the single assumed morphological mould. This was also the 

 method long pursued in the comparative morphology of the Angio- 

 spermic flower. 



4. The use of single characters for purposes of comparison. This 



