64 GYMNOGRAMMOID FERNS [cii. 



their sporangia. This section would comprise L/avea, Cryptograinme, Onychium, 

 Ceratopteris, and Jamesonia. 



II. A second section would centre round Gynutogmimne 2ind Hemionitis, 

 together with such minor genera as Pterozonium, Syngravime, Anogramme, 

 Coniogramme, Gymnopteris, Ceropteris, and Trismeria: all of which have 

 from time to time been included in the old genus Gymnograimne Desv., i8i i. 

 Though these distinct genera may be upheld as now defined, there is no 

 reason to doubt their near affinity to the central genera named. It is necessary, 

 however, to remark the exclusion of the genera Ceterach Adanson, 1763, and 

 Pleiirosorus Fee, 1850; for they are now regarded as partially or completely 

 non-indusiate Asplenieae: also the position o{ Aspleniopsis Mett. and Kuhn 

 must be held in suspense (see Nattirl. Pflanzenfam. I, 4, p. 272, Fig. 145, A). 



III. A third section would consist of the Adiantinae, represented by 

 Adiantum itself. 



IV. A fourth section includes specialised xerophytic types, such as Pel/aea, 

 Doryopteris, Cheilanthes, and NotJiolacmx, together with Saffordia and 

 TracJiyptcris. 



The central features that all these Ferns have in common are that the 

 sorus is superficial in origin, never actually marginal, and that there is not any 

 vestige of an inner indusium. Nevertheless the leaf-margin may be variously 

 developed and recurved, so as to give protection to the sporangia. The 

 receptacle often extends a considerable distance along the underlying veins, 

 while in extreme instances the insertion of the sporangia may not merely 

 extend along them, but also spread laterally from them so as to cover the 

 leaf-surface, in the manner known as Acrostichoid. It will be realised that 

 these features are not in themselves very distinctive, and that they fall short 

 of being clearly diagnostic. If the original position of the sporangia in Ferns 

 generally was, as believed, distal or marginal, a transit to the surface might 

 appear in any phylum where the leaf-area became extended (see Vol. I, 

 p. 225). Thus along a number of distinct phyletic lines the Gymnogrammoid 

 character may have been acquired. The natural inference will then be that 

 Ferns showing that character may belong to several distinct lines of descent, 

 and not necessarily be phyletically related to one another at all. Accordingly 

 we must be prepared for some degree of segregation, and possibly for the 

 recognition of affinity of one of these sections, or even of some single genus, 

 to one phyletic source; while others may be referable to some quite distinct 

 source. These preliminary remarks form a necessary introduction to the 

 description of the more prominent genera grouped as already suggested: they 

 leave the question of the phyletic unity of the Gymnogrammoid Ferns 

 open for discussion after the facts have been disclosed. 



