XLix] DIPTEROID DERIVATIVES 281 



the Dipteroid derivatives comprise a considerable body of P'erns, whether 

 fossil or living: and that the series has pursued a course of its own, inde- 

 pendent of other phx'la, from the Triassic Period to the present day. It 

 originated from a source related to other early superficial Simplices, and 

 ends in types which have been referred to the old comprehensive genera, 

 Acrostichuin and Polypoditim. Such a progression is thus homoplastic with 

 others that passed through a similar though phyletically distinct history. 



There remain a considerable number of genera of Ferns commonly of 

 small size, with their unprotected sori more or less distinctly superficial, and 

 often elongated, following the course of the veins. These genera have been 

 associated habitually with Gyuuiogrammc, and in point of fact many of them 

 have been from time to time included in that old and comprehensive genus 

 (Chapter xxxiv). They were grouped by Diels with the Pterideae, as 

 Pterideae-Gymnogramminae: but this presumes for them a secondary origin 

 by abortion from an indusioid type, of which there is no evidence: they all 

 appear to have had unprotected sori from the first. In the Osmundaceae, and 

 certain Schizaeaceae, together with Plagiogyria we see relatively primitive 

 Ferns which had non-indusiate sori, and it is in relation with these that 

 their natural place may be sought. 



The Gymnogrammoid Ferns fall into four natural groups: (i) those which 

 have primitive characters such as Llavea and Cryptograimne: (ii) those 

 which centre round Gyninogramine and Hemionitis: (iii) the well-marked 

 genus Adianiuni: and (iv) a group of specialised xerophytic types associated 

 with Cheilanthes and Notholaena. There is no sufficient ground for assuming 

 that these four groups were of common origin, for the features on which 

 they are associated together are negative rather than positive. Llavea and 

 Cryptograimne have usually been held as related to Plagiogyria: and 

 this, with a further reference to Todea, may be held as a probable source 

 for them. Perhaps also for Ceratopteris and Jamesojiia, though with less 

 certainty of reference as between an Osmundaceous and a Schizaeoid origin 

 (Chapter XXXIX, p. 72). 



The second group presents a more difficult problem : they no doubt form 

 a coherent body of genera, and are probably related to the first group: but at 

 present their definite reference must be left uncertain. The same may be said 

 of the genus Adiantiim which forms the third group (Chapter XXXIX, p. 78). 



The fourth, or Cheilanthoid group, with its sparse sporangia borne close 

 to the protective margin, and the habit so often suggestive of MoJiria, has 

 long been compared with the Schizaeaceae, a comparison which detailed 

 study of both generations has strengthened. Though this falls short of actual 

 demonstration of affinity, there is reasonable probability that Notholaeria 

 and Cheilanthes are Schizaeoid derivatives from the type of Mohria, their 

 sporangia having passed from the margin to the surface of the blade, in a 



