

18 THE BRITISH FERNS. 
to as the ultimate pinnules or segments, according as they may be 
either entirely or only partially separated. 
It will now be seen, and it is a point important to observe, that 
the costa and rachis of a simple frond, are not the same parts as 
those of a pinnate, or of a bipinnate frond, but that the application 
of the terms changes according to the degree in which the frond is 
divided. The costa is the principal rib of the simple frond, or of 
the last series of distinctly formed leaflets in divided fronds; while 
the rachis in a divided frond is the part answering to the costa in 
the simple. frond, that is to say, the direct continuation of the stipes 
through the lamina. The position of the costa once settled, the 
first branches from it are always the veins, the branches from these 
are the venules, and the branches from the latter, the veinlets. The 
foregoing statements respecting the application of the term costa, 
have been made to apply to those fronds which are either simple or 
distinctly divided into separate leaflets. It, has been already men- 
tioned, however, that there is another mode of division intermediate 
in character between these, called pinnatifid, bipinnatifid, tripinna- 
tifid, &e., the divisions in these cases not being separate leaflets, but 
lobes, more or less deeply separated, of the scolloped margin. This 
does not affect the position of the costa, unless the lobes are very 
shallow indeed, so that in general the midrib of the lobe would 
also be called the costa. — 
The venation of the frond, which has now been shown to mean the 
costa and its ramifications, presents great variety in different species, 
but it is, in a general way, quite constant in the same species. 
There are a few instances known, in which, for example, Ferns 
which usually have the veins free, have them here and there 
running together in particular fronds or parts of the fronds, but 
these are to be regarded as mere exceptions which prove the general 
rule, and are not often liable to be misunderstood. Indeed this 
inconstancy most generally happens in cases where some disturbance 
