Salmon . — On the Genus Fissidens. 107 
dorsal and ventral surfaces of the nerve, that is to say, they 
never occur laterally. Therefore the nerve-structure, as shown 
in Figs. 2 and 3, would indicate by the two £ stereid 5 bands 
(b, b f ) the upper and under surfaces of the leaf. But these 
sections are taken at a point where, according to Robert 
Brown’s theory, the leaf bears an appendage at the back 
of the nerve ; and to see the normal nerve-structure, we must 
select a species with leaves in which the appendage — the 
inferior lamina — dies out before reaching the stem. Such 
a species is F. fascicidatus, Hornsch., and Fig. 5 shows 
a section of the leaf just above its insertion. Here we see 
very plainly that the nerve-structure is that of an ordinary 
sheathing leaf, with the ‘ stereid ’ bands (a) ( b ) marking 
respectively the upper and under surfaces of the leaf. If 
we follow, by means of serial sections, the nerve-structure 
as we ascend the leaf, we shall see exactly how the arrange- 
ment shown at Figs. 2 and 3 has arisen. At Fig. 6 we 
find the first appearance of cells belonging to the inferior 
lamina, and at once the nerve-structure becomes altered. 
The stereid band marking the back of the leaf becomes 
divided by unthickened cells (the ‘ conjunctivae ’ of Lorentz) 
connecting the inferior lamina with the large cells (‘ Deuter ’) 
of the nerve proper. Fig. 7 shows the inferior lamina more 
developed, the two halves of the dorsal stereid band pressed 
more to the side, and the ventral stereid band beginning to 
die out. Fig. 8 corresponds to Figs. 2 and 3, and we can 
compare them. In Fig. 8 we see that the groups (b, b') 
belong to a single dorsal band and that the cells (a) are 
the remains of a ventral stereid band, and mark the inner 
surface of the leaf : the rows of large cells connecting the 
laminae consist partly of the ‘ Deuter ’ of the original nerve- 
structure, and partly of cells placing the inferior lamina 
in connexion with these. In Figs. 2 and 3 the groups (b, b') 
are similarly the divided halves of a dorsal band, and do 
not indicate the upper and under surfaces of the leaf ; there 
is here no trace of a ventral stereid band, but we can see 
clearly that the upper surface of the leaf lies towards (a). 
