326 GENERAL REVIEW OF THE PRIMITIVE FERNS [CH. 



slide of the sorus from the margin to the surface of the broadening leaf. But 

 this did not happen once for all. In certain Ferns the transition was effected 

 early, in others late. Hence the distinction is not an absolute one. But as 

 the sori of the Schizaeaceae themselves are always marginal in primary origin, 

 and those of the Gleicheniaceae are as constantly superficial, the distinction 

 in respect of this character dates from Palaeozoic time. The families closely 

 related to these two ancient types show for the most part a similar constancy 

 in position of their sori. Thus the distinction between the ancient Marginales 

 and Superficiales must be held as a real one and phyletically early. 



The families that group themselves naturally with the Schizaeaceae on 

 the sum of their characters, including that of soral position, are the Hymen- 

 ophyllaceae, Loxsomaceaeand Dicksoniaceae (including theThyrsopterideae, 

 Dicksonieae, and Dennstaedtiinae), and they may be held as relatively 

 primitive Marginales. Those that group themselves naturally with the 

 Gleicheniaceae on the sum of their characters, including that of soral position, 

 are the Protocyatheaceae {Lophosoria and Metaxyd), the Cyatheaceae (excl. 

 Dicksonieae), and the Dipteridaceae. The curiously isolated family of the 

 Plagiogyriaceae, represented only by the living genus Plagiogyria, is difficult 

 to place. It certainly is a relatively primitive type, and may very possibly 

 represent a direct offshoot from the Osmundaceae, of the type of Todea, with 

 which it shares the upright habit, open venation, certain points of anatomical 

 structure, and the superficial sorus. All these Ferns may be held as relatively 

 primitive Superficiales. 



This grouping separates the Dicksoniaceae from the Cyatheaceae, 

 a separation which is long overdue. Their grouping collectively under the 

 title of the Cyatheaceae was in the first instance based upon their sharing 

 the dendroid habit. Strong indications of their real distinctness lie not 

 only in the constant difference of their soral position, but also in the two- 

 sided segmentation, with a four-rowed stalk, seen in the sporangia of the 

 Cyatheaceae, a point which they share with the Dipterids, but which is 

 absent so far as yet observed in the Dicksoniaceae. Further, there is the 

 difference in the indusium, which is absent in Alsophila as it is also in 

 Lophosoria and Gleichenia\ but constantly present in two-lipped form in the 

 Dicksoniaceae. Again the consistent presence of dermal hairs in the 

 Dicksoniaceae, as against the broad scales of almost all the Cyatheaceae, is 

 a strong indication of distinctness. Finally, the near relation of LopJiosoi'ia 

 on the one hand to Gleicheiiia and on the other to Alsophila, all of them 

 having superficial sori, is strong evidence of a transition from the simple 

 superficial type to the gradate without any relation to the marginal series 

 at all. If it had not been for the common feature of a dendroid habit it is 

 improbable that Ferns differing in so many material features would have 

 been classed together. The differences leave little doubt of the propriety 



