522 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vox. VIII. 
Occasionally leaf-gaps occur in Osmunda cinnamomea through which 
internal and external endodermes connect and even pith and cortex 
(Fig. 11). But the typical leaf-gap is occupied largely by parenchyma 
which is not penetrated by the endodermis, but into which a cone (as it 
appears in transverse sections) of the external phloem projects more 
or less deeply, and outwards into which the internal endodermis often 
tends to run (Fig. 11). In the typical leaf-gap, too, a blind pocket extends 
from the inner end of the ray down into the xylem, or—to explain in 
another way—on examining an ascending series of transverse sections 
of the stem, one finds nests of parenchyma making their appearance in 
the strands of xylem, which on being traced upwards are found to be 
continuous with the parenchyma of the medullary rays. It is worthy 
of observation that these nodal pockets, as they may be called, are just 
centrad of and in direct contact with the protoxylem elements of the 
strands in every case. Similar pockets are quite commonly absent in 
the early medullated stages of the seedling of Osmunda cinnamomea, 
and now and then in the adult, but they constitute such a characteristic 
feature of the Osmundacee, living and fossil, (cp. with the nodal pockets - 
Gleichenia pectinata, and the Gleicheniacez) that a study of the phenomena 
connected with them may prove helpful in clearing up some phylogenetic 
tangles. 
The first indication of a parenchyma pocket in the stele of Osmunda 
cinnamomea is represented in Fig. 15a by three cells in the xylem. These 
are continuous with the cells occupying the bay in 15b, and the pith in 
I5c. These sections were taken in the neighbourhood of the sixth node 
of specimen A, and as has been already stated the stele returns to the 
protostelic type before the next node is reached. In Fig. 14 (from speci- 
men E) another pocket of three cells is represented—in this case it is the 
end of a ‘‘bay”’ which has been closed off by one or two tracheids as is 
shown. ‘This stele likewise resumes the protostelic type farther up the 
stem, but neither here nor in the case represented by Fig. 15 can the 
pocket be affirmed to be homologous with the nodal pockets of the adult, 
But the remaining instances cited may reasonably be regarded as such, 
Fig. 16 represents a section higher up in specimen E, and as is evident 
there is a narrow (though not permanent) pith. A leaf-trace is just 
separating, its upper edge has already broken away, and the position 
of the lower edge is marked by a small x in one of the tracheids. The 
‘‘pocket”’ at this level is wider than the pith and has been continuous 
with the latter from the first. In the same figure, there may readily 
be detected two contiguous parenchyma cells, one with a nucleus, in 
the bulkiest part of the xylem—these constitute the bottom of the suc- 
