532 TRANSACTIONS OF THE CANADIAN INSTITUTE. [Vou. VIII. 
external endodermis and cortex through leaf-gaps, but rarely, and so far 
instances have been observed in adult stems only. From their first ap- 
pearance, however, they very commonly extend outwards into the leaf- 
gaps. 
3. Branching is not a seedling character. 
4. Internal phloem has been found in unbranched adult plants. 
This. fact, and the absence of branching in the seedling are believed .to 
indicate that there are no adequate grounds for regarding internal phloem 
and internal endodermis in the Osmundacee as having ‘‘intruded”’ 
through branch-gaps. ; 
5. The parenchyma pockets extending down from the medullary 
rays into the xylem are sometimes occupied by an isolated portion, as 
viewed in transverse sections, of the internal endodermis. Often in the 
seedling and occasionally in the adult they are represented merely by 
grooves on the inner face of the cylinder of xylem. The origin of these 
pockets maybe accounted for by the tendency of the xylem to encroach on 
the axial tissues, thus closing around the lower ends of the oblique medullary 
rays. This tendency is seen in the occasional closure of the inner entrance 
to the medullary ray, and in the existence of larger or smaller isolated 
strands of xylem on the pith side. ‘Tracheids may even occur in the pith, 
in which case they are surrounded by a ring of endodermal cells. ‘Thus 
the nodal pockets of the Osmundacee may furnish important evidence 
of an evolutionary trend in this family towards cladosiphony. 
The ‘‘cladosiphony”’ of Osmundites Dunlopt, etc., is not infrequently 
paralleled in plants of Osmunda cinnamomea that have been growing 
under adverse conditions. 
6. The stele of the existing Osmundacee is regarded as a reduced 
amphiphloic siphonostele, and hence the homologies of the pith, and the 
origin of the siphonostele as concordant with the stelar theory of Jeffrey 
in its application to the Filicinee. 
University of Toronto, June, 1909. 
