94 GYMNOGRAMMOID FERNS [CH. 



than the rest, for in them the sporangia are standardised after the usual 

 Leptosporangiate model, and are disposed for the most part in linear sori 

 along the veins, but stopping short of their actual endings. These Ferns 

 possess dermal scales, and a vascular system derived from solenostely, but 

 with an occasional tendency to perforation, with a binary leaf-trace (Vol. i. 

 Fig. 1 60). There is also reticulate venation in some of them, with other 

 signs of advance including a tendency to Acrostichoid development. It 

 would be easy to see in them the results of specialised advance from such 

 types as have been above held as primitive: but the phyletic lines have 

 not been exactly traced, so as to give such a suggestion any degree of 

 certainty. 



The Maidenhair Ferns {Adiautiwi) may with greater confidence be 

 referred to some central Gymnogrammoid or Cheilanthoid source : but here 

 the linear sori are borne upon the distal region of the veins, and the 

 marginal lappet that bears them has become sharply reflexed. This arrange- 

 ment was clearly successful, as shown by the existence of nearly 200 

 species. But the genus offers little sign of advancing specialisation beyond 

 occasional vein-fusions {Hezvardid), and a tendency to an Acrostichoid state 

 {Adiantelld). There is, however, a structural detail of the sporophyte that 

 acquires importance in relation to the Vittarieae; that is, the existence of 

 "spicular cells" in the epidermis. In the gametophyte also the collenchy- 

 matous thickening of the walls of the prothallial cells presents a feature 

 that has a special comparative importance in relation to the Schizaeaceae. 



Of all the Gymnogrammoid Ferns it is the Cheilanthoid group that 

 presents the highest phyletic interest. Robert Brown, in founding the genus 

 Notholaena, speaks of "sori marginales" {Prod. Fl. Nov. HolL p. 145, 18 10): 

 the very name CJieilantJies introduced by Swartz (1806) conveys the same 

 for that genus. Thus the earliest writers on these genera held their sori as 

 actually marginal. On the other hand. Hooker {Syn. Fil. p. 436) speaks of 

 MoJu'ia cajfroruni as combining the capsules of the sub-order (Schizaeaceae) 

 with the habit of Cheilanthes. Thus early the habit-similarity was recognised 

 between the Schizaeaceae and the Cheilanthoid Ferns, while a marginal 

 sorus was attributed to the latter. These historical notes form a natural 

 preface to their comparison in the light of details later acquired. 



First comes the question of fact, as to the relation of the sporangia to 

 the margin. Prantl {Sdiizaeaceen, 1881) demonstrated most fully that the 

 sporangia of MoJiria arise from marginal cells (Fig. 645), but that they were 

 early displaced to the lower surface by the growth of an indusial false 

 margin : the later result of this is seen in Vol. II, Fig. 450. But observations 

 on Notholaena, in which of all Cheilanthoid Ferns the sporangia are most 

 nearly marginal, have shown that in point of fact they arise superficially, 

 though the sporangium nearest to the marginal cell may be only one segment 



