XVII] ORIGIN OF LEAF 341 



with apical growth. The fact thus brought into the comparison is not quoted 

 as any indication that Tmesipteris represents an ancestral form for Ferns. 

 It is cited to show that the 

 condition required by our in- 

 duction from the develop- 

 mental facts has now been 

 actually observed in a primi- 

 tive type of living sporophyte. 

 Thus for the first time the 

 suggestion made ini 884 takes 

 its foundation upon a basis 

 of observed fact. The hypo- 

 thesis may now be accepted 

 as passing from the region of 

 mere surmise to the arena of 

 scientific theory. Once the 



distinction between axis and Fig. .^09. Young sporophyte of Tmesipteris still attached to 



leaf is attained the further ^^^ prothallus, cut in longitudinal section, and showing 



J . . r 't- , ^ '^^'° equally developed shoot-regions. (After HoUoway.) 

 derivation of a Fern-leaf, as 



seen in its simpler forms, will follow readily by reduction of the radial to a 

 bifacial type, and by sympodial development and webbing of the dichoto- 

 mous twigs to form the pinnae, along lines traced in Chapter V. Finally, if 

 such a distal position of the sporangia as is seen in the early Devonian forms 

 were retained throughout the progress of these changes, they would then 

 appear terminal on the vascular strands, as they actually are in Stanropteris, 

 and in Botrychium (compare Lignier, 305, 308). 



The origin of the root-system is not clear. The distended protocorm, so 

 prominently seen in some of the Devonian fossils, does not appear in living 

 Ferns. It may never have been a feature in them. It may be suggested that 

 the prototype of the root is to be seen in the branched rhizome of such 

 a type as Asteroxylon. But of this there is no direct evidence, and the question 

 of its origin for the Filicales must be left undecided. There is, however, 

 some reason to think that the most primitive Ferns may have been rootless, 

 as the Psilophytales are, a view v^'hich readily accords with the adventitious 

 character of the later roots, and the appearance of the first root in the embryos 

 of some primitive Ferns in a manner that clearly suggests an accessory organ. 



Passing from these comparisons in respect of the sporophyte to those 

 of the gametophyte the ground is less satisfactory, partly because the features 

 of the prothallus are less pronounced and more variable, partly because fossil 

 prothalli are almost unknown, and comparison is thus thrown back solely 

 upon those of modern Ferns. In Chapter XV it was concluded that a fila- 

 mentous structure probably underlies all gametophyte-development in Ferns. 



