776 MR R. KIDSTON AND MR D. T. GWYNNE-VAUGHAN ON 



appears to have been the case in most of the other Pteridophyte dictyosteles and 

 solenosteles. 



This idea leads us to regard Osmunda cinnamomea as illustrating one of the simpler 

 stages in an ascending series of cladosiphonic structures that have culminated in the 

 complexity of Osmundites skidegatensis, and certain points in the anatomy of the former 

 plant seem distinctly in favour of this view. For instance, the internal phloem doea 

 not occur in all regions of the stem, but is more or less closely confined to the neighbour- 

 hood of the points of branching. Even at these points it is sometimes wanting, as in 

 those cases when the stele branches by a simple median constriction without forming a 

 gap in the xylem ring. Indeed, it seems that it is never present under such conditions ; 

 but existing observations on this point are not conclusive. Again, whenever the internal 

 phloem actually is present it is invariably continuous with the external phloem through 

 a gap caused by the branching of the stele ; and further, the amount of internal phloem 

 present is closely related to the extent of the opening in the stele. The significance of 

 these facts is accentuated by the phenomena that we observed in a case of branching in 

 Todea barbara. Here the xylem ring of the stele opened up in the sinus between the 

 two branches, and the external phloem subsided through the gap for a considerable 

 distance into the medulla of the main axis below the branching. It should be mentioned 

 that in this case neither the porose layer nor the endodermis was decurrent. It is 

 possible, however, that further developments on this line might be met with if other 

 cases of branching could be obtained. 



If it were once conceded that the internal phloem obtained admission in this clado- 

 siphonic manner, it is easy to conceive of the subsequent changes that would result in the 

 structure of Osmundites skidegatensis. Concurrently with the distension of the stele 

 and the widening of the leaf-gaps incidental to the departure of such large leaf-traces as 

 are possessed by this plant, the phloem would tend to project more and more deeply into 

 the medullary rays both from the inside and from the outside. In fact, according 

 to Jeffrey and Faull it is already beginning to do so in Osmunda cinnamomea. Then 

 the internal and external phloems would meet so as to form a phloem ray, and later the 

 central ground-tissue would project into the median region of this phloem, separating it 

 into two layers lining the sides of the leaf-gap. The two strips of parenchyma lying 

 between these layers of phloem and the xylem now represent all that is left of the original 

 medullary ray tissue. A still further outward extension of the internal ground-tissue 

 would cause the complete disruption of the peripheral tissues of the stele at the level 

 of the departing leaf-trace (cf. text fig. 1), and the structure of Osmundites skidegatensis 

 would now be attained. Regarded in this light, the continuity of the internal ground- 

 tissue with the external is a secondary and not a primary phenomenon. 



The fact that an internal endodermis is always present in Osmunda cinnamomea, 

 even in those parts of the stem where no phloem is to be found, may possibly be 

 advanced as an objection to this theory. However, there is, in the first place, no 

 necessity whatever to assume that the internal endodermis actually originated m 



