POLYPODIACE^E. 329 



hereditary 1} rather than posterior, and are subject to the least change in the 

 course of the phylogcnical development of species. It is, therefore, one of the 

 most reliable characters for the determination of the kinship of genera. 



Let me now briefly state how far Prosaptia agrees with Davallia in 

 respect of its vegetative organs. The rhizomes of the latter are long-creeping 

 with remotely arranged fronds, while those of the former are short-creeping 

 with densely arranged fronds. The scales of Davallia are composed of cells 

 with very thin reddish walls arranged in one layer with no trace of bristles, 

 and are totally different from what we have observed in Prosaptia or Poly- 

 podium obliquatum. The hairy stipes so peculiar to the latter two have never 

 been, found in Davallia. So far as the vegetative organs are concerned, 

 Prosaptia is not congeneric with Davallia. 



Now let me consider more precisely a comparison of the same ferns in 

 respect of their fructification. My opinion is that the fructification of Prosap- 

 tia is altogether the same as that of Polypodium in its origin, but totally 

 different from that of Davallia in its beginning, as I shall show later on. 

 Before I go into details, I must pause to consider what the fructification of 

 Polypodium obliquatum or of its allied species is like. In the full grown 

 form, the sori (Fig. 137, g) of this fern look very different from the ordinary 

 sori of a normal Polypodium. Therefore, it was regarded by FEE as repre- 

 senting the special genus, Cryptosorus-^ . The sorus is located in a pouch -like 

 cavity immersed in the tissue of the fronds on the under surface, which cavity 

 has an orifice beset with bristles. According to the author, the sori develop in 

 the following manner : they are in the beginning formed under the epidermis 

 (" cuticule ") from which, when fully developed, at length break out ; the slits 

 in the spidermis are the orifices of the pouch-like cavities. In order to assure 

 myself of what is stated by FEE, I looked in my very rich collections of the 

 same fern, and was so fortunate as to find that the fern is of a nature of 

 indefinite growth, representing several stages of the sorus-development in one 

 and the same frond. On the basal portions of the frond, I found full-grown 



( CHKISTENSEN, C. On the natural classification of the species of Dryopteris, p. 75. 

 ISOILEREDER, H. Systematische Anatomic cler Dikotyleclonen, p. 937. 

 (2) FEE, A.L.A. Lc.p. 231. 



