ON GENERA AND SPECIES. 63 
tute an important part in defining genera; but in many 
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 characterised 
by having linear naked sori; but by adhering to this 
simple character alone, species of very different habit 
would be included, as Gymnogramma, Stegnogramma, Loxo- 
gramma, Plewrogramma, Sellegua, and others, which, as 
already stated, has to a considerable extent been so done in 
the "Species Filicum." The 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 Pleopeltis, Phymatodes, Pleuridium, Drynaria, 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 punctiform, and produced in a single row between 
each pair 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 Gymnogramma, as characterised in the “ Species 
. Filicum.” These examples show that the form of the sn 
alone cannot be strictly depended upon for defining genera. 
In previous remarks I have stated that the mode in 
which the vascular structure traverses the fronds (vena- 
tion) affords important data in assisting to define the limits : 
of genera, : 
