ON GENERA AND SPECIES. 63 



tute au important part in defining genera ; but in manj' 

 cases if the characters derived from the form are strictly 

 adhered to, species possessing very different appearances 

 are brought together. For instance, the small grass-like 

 Ferns of the genus Grammitis of Swartz are charaoterised 

 by having- linear naked sori ; but by adhering to this 

 simple character alone, species of very different habit 

 would be included, as Gyninogramraa, Stegnogramma, Lu.ro- 

 (jramma, Pleurogramma, Sellegua, and others, wlwch, as 

 already stated, has to a considerable extent been so done in 

 the " Species Filicum." Tho sori are also liable to vary in 

 form and size in different specimens of the same species, or 

 even on the same, or different fronds of the same plant. 

 This usually happens by the contiguous punctiform recep- 

 tacles uniting- laterally, when they assume on oval, oblong, 

 or linear form. Examples of this may be found in the 

 genera Pleopelfis, Plujinatodes, I'leuridium, Dryvaria, Aspi- 

 dium, and others. A remarkable instance of this occurs in 

 the splendid Fern Drynaria coronans {Polypodium coronans, 

 Wall.), a species with large pinnatifid fronds, which in the 

 presumed normal state as generally observed, has the sori 

 distinctly puQctiform, and produced in a single row between 

 each jjair of the primary veins. In some fronds, however, 

 the sori are so completely united in a line as to form a true 

 linear sorus, and hence not only agree perfectly with the 

 character of Grammitis, as defined by Swartz, but also with 

 the genus Gi/mnogram>iia, as characterised in the " Species 

 Filicum." These examples show that the form of the sori 

 alone cannot be strictly depended upon for defining genera. 

 In previous remarks I have stated that the mode in 

 wdiich the vascular structure traverses tho fronds (vena- 

 tion) affords important data in assisting to define the limits 

 of p-enera. 



