218 CONCLUSION 



From this central type presumably several lines diverged, of which only a few frag- 

 ments persist. One of these is seen in the different forms of Hot r\c In urn, to which 

 Helminthostachys is probably not very remotely allied. The whole of this series is 

 characterized by a subterranean gametophyte, and a more or less saprophytic habit 

 of the sporophyte is indicated by the development of a mycorrhiza in the roots. 



The Marattiaceae, as they now exist, probably do not represent a single un- 

 broken line of descent, and show strong evidences of a possible multiple derivation 

 from the earlier stock. The point of contact with the Ophioglossaceae is probably 

 in the neighborhood of Helminthostachys, which, on the whole, is more like the 

 Marattiaceae than are the other Ophioglossaceae; but it is improbable that the solid 

 synangium, such as characterizes most of the Marattiaceae, was derived from a 

 group of separated sporangia like those in Botrychium or Helminthostachys, and it is 

 more likely that they arose from a structure which resembled the spike of Ophiu- 

 glossum. Angiopteris is, with little question, the most specialized of the living 

 Marattiaceae and has probably departed further from the ancestral type than any 

 of the other forms, while Kaulfussia, on the other hand, is probably the most 

 primitive. On the whole the Marattiaceae are nearer to the leptosporangiate ferns 

 than the Ophioglossaceae are, and it is likely that the Leptosporangiates are directly 

 descended from some ancient fern forms, allied to the Marattiacea?, but differing 

 from any of the existing types. 



