io6 Salmon . — On the Genus F, is si dens . 
Brown’s explanation seems sufficiently good when treating 
of the species of Fissidens that possess a nerve, but that 
it cannot be applied to those without a nerve. 
Lindberg’s theory has been supported by Braithwaite ( 5 ), 
who makes the following remarks on the different theories : — 
f One of these [theories] was that the leaf to a certain extent 
was split vertically to embrace the stem ; but this is not 
tenable, as each half of the split portion is of equal thickness 
to the rest of the leaf. Another view which has met with 
general acceptance is that the double portion alone is the 
true leaf, . . . and all the rest is an outgrowth from it. . . . 
However plausible this view may appear, it does not satisfy 
us, for there are species in which the duplicate part is nearly 
or altogether wanting, e.g. F. dealbatus. The most rational 
explanation seems to be this, that the additional lobule is 
of the nature of a stipule, arising on the opposite side of 
the stem, which has become adnate to the nerve by the 
whole lower margin, the upper margin being free and parallel 
to the corresponding margin of the leaf. 5 
Passing these three theories in review, we notice that 
according to Robert Brown’s theory alone the true leaf 
(formed of the vaginant laminae) is considered to be hori- 
zontally inserted ; in the two other theories the whole ex- 
pansion is considered to be the leaf (either split on one side, 
or united to a stipule) with a vertical insertion. If, therefore, 
the internal structure of the leaf can decide the question of 
insertion, it will give important evidence in deciding between 
the theories. Such proof is, I think, furnished by the nerve- 
structure of certain species. 
The usual type of nerve-structure in Fissidens , as shown 
in a transverse section near the base of the leaf, is represented 
at Figs. 2 and 3. We find two well-developed ‘stereid’ 
bands ( b , b '), separated by a single or double row of large 
thin-walled cells. If we refer to the works of Lorentz (6, &c.) 
on the structure of the nerve in Mosses, we find that where, 
in the many hundred drawings of nerve-sections which he 
gives, two ‘ stereid ’ bands exist, they invariably mark the 
