248 HYMENOPHYLLACEAE [CH. 



Comparison 



From what has been stated above there is no reason to regard the 

 Hymenophyllaceae as anything more than a relatively primitive family of 

 Ferns, showing progressive degrees of specialisation in relation to a hygro- 

 philous habit. This reaches an extreme in the smaller species of Trichomanes, 

 and it finds its expression in their minute size, and simple structure: in the 

 substitution of "pseudo-veins" for the normal venation: in the stele some- 

 times having no xylem at all: and in the absence of true roots in the 

 smallest species. It is also apparent in the greatly reduced spore-output 

 per sporangium. It may be a question whether or not the moist habit may 

 not also be answerable for the simplicity of the gametophyte. If such 

 characters be taken as results of hygrophytic specialisation, the problem 

 will be which types of the family are to be held as primitive, for it is through 

 these that the probable relationship of the family may best be traced. In 

 each of the two genera, which have probably undergone independent 

 specialisation, certain larger species appear to be marked out. Of these 

 Tiichomanes reniforme has always been regarded as a peculiar and out- 

 standing type. It was isolated generically as Cardiomanes by Presl. Prantl 

 draws attention to its possessing the simplest leaf-architecture, with the sori 

 borne on the ends of its forked veins (Fig. 504, ^). Moreover its spore-output 

 is high (256), as it is also in the larger species of Hymenophyllum. In its 

 anatomy also it approaches them. These characters suggest that T. reniforme, 

 together with some large species of HymenopJiylluin, such as H. dilatatum, 

 may be held as taking a central position in the Family, from which divergent 

 lines of specialisation may have radiated. H. dilatatum is confirmed in that 

 position by its Sphenopteris-typQ of leaf-construction, its well developed 

 vascular structure, and fairly high spore-output (128): but T. reniforme has 

 its leaf fully webbed. 



Taking then the characters of the sporophyte, as seen respectively in these 

 two central types, they may be compared with those of other Filicales, The 

 creeping habit is already seen in the Schizaeaceae and Gleicheniaceae, 

 which are also protostelic. But the peculiar structure of the stele in the 

 species named finds its nearest correlative in the Botryopterideae, and 

 especially in AnacJioropteris. Allowing for the differences which follow on 

 an upright habit and radial construction in this fossil and the creeping habit 

 of most Hymenophyllaceae, the structural resemblance is very close: and 

 with it goes a striking similarity in the structure and insertion of the leaf-trace, 

 and in the mode of supply to the axillary buds. H. dilatatum and T. reni- 

 forme are both species with a lamina several layers of cells in thickness, and 

 occasionally showing alternate segmentation of the marginal cells, as in 

 Leptosporangiate Ferns. It may be a question what interpretation these 



