[kirsca] CERTAIN STRUCTURES IN THE PTERIDOPHYTES 397 
thus block the canal at a comparatively early stage of development, and 
that is what has probably occurred in the stipes of Osmunda that were 
examined. In this case the thyloses have blocked the canal, but have 
not yet had time to thicken their walls very much. 
It is these thylosal cells that De Bary (3, 347) speaks of in dis- 
cussing the vascular bundle of the petiole of Osmunda. “On the con- 
cave side it is parenchymatous, and distinguished by 10-12 small groups 
of conspicuously wide cells, which still need investigation. It is by no 
means surprising that he failed to interpret them, since he only saw the 
last stage of the structure and not its early development. 
Of the Marattiaceæ, Angiopteris evecta was examined, stipes of 
this form being available. 

Fig. 26.—Angiopteris evecta, racais. Sbcwing canal, C. with thyloses, Th., 
growing into it. X 245. 
On examining very thin sections of the stipe, there were seen to be 
large, irregular areas of clear, thin-walled cells next to the protoxylem 
in many of the bundles (Photo 20). In other bundles there were 
cavities adjacent to the protoxylem, with cells of a similar nature to the 
above jutting into them (Fig. 26). These cells are the thyloses men. 
tioned above for the other forms, and present identical characteristics. 
Remains of the protoxylem in the form of rings and spirals, as well as 
drawn out spirals which appear like glistening threads, are frequently 
found in these areas. The thyloses are very clear of contents, but often 
exhibit prominent nuclei. | 
