Boodle. — Anatomy of the Gleicheniaceae. 719 
observed in most of the endodermal cells of the petiole, 
both in the type and in the form originally separated as 
G. rupestris . These bodies, which are just like those in the 
endodermis of the stem of G. flabellata^ were not seen in the 
petiole of any other species besides that mentioned. 
In the bundle of G. pedatis the endodermis is only slightly 
incurved, so there is room for a mass of sclerotic pericycle, 
which is present here as in some Eugleichenias , between the 
endodermis and the cavity-parenchyma accompanying the 
median protoxylem. The cavity-parenchyma, as in several 
other Mertensias , forms a conspicuous group of lignified cells. 
In the large form of G. longissima previously mentioned 
the fibres of the bundle are specially numerous. The bundle 
is of large size, and the fibres extend practically all the way 
round the inner face of the xylem. Opposite the terminal 
protoxylems they form a nearly continuous layer, and appear 
to belong to the phloem, elsewhere they are more scattered 
and perhaps represent pericycle. 
The petiole of G. cryptocarpa is remarkable, as it contains 
three bundles instead of one. They are not widely separated, 
and if pushed into contact would form a typical C-shaped 
bundle with the usual orientation. There is one long slightly 
curved bundle, and two shorter sharply curved ones, repre- 
senting the long side and the two hooked ends of the C 
respectively. As a single leaf-trace passes off from the stele 
to supply the leaf, and is continued into the base of the 
petiole as an arched bundle of the usual form for Mertensia , 
the peculiar structure described above is to be explained as 
due to a very early separation of the bundles for the first two 
pinnae 1 . This division of the bundle generally takes place 
quite a short distance below the pinnae, but here it had 
already occurred at fourteen inches below the first pinnae. 
In G. flageltaris and some other species each incurved end 
of the xylem-arch does not terminate in a hook, but in a knob, 
in the middle of which a protoxylem-group is placed. This 
feature may be more or less resfficted* tb the upper part of 
1 The same thing appears to take place in G.Jlagellaris (Poirault, ’93, p. 176). 
