262 
A. G. Tansley. 
exists in the neighbourhood of the forks of the stem, is more 
primitive than the typical structure of, for instance, 0. regalis, and 
represents a degraded form of the “amphiphloic siphonostele,” 
which is supposed to have been the ancestral type. From the type 
of O. cinnamomea such a form as O. regalis is supposed to have 
arisen by the complete disappearance of the vestigial internal 
phloem and of the internal endodermis, and the complete shutting 
off of the pith from the cortex. The discovery of 0. skidegatensis 
would seem to have furnished these authors with their hypothetical 
ancestral type from which to derive O. cinnamomea. 
The arguments that appear to me to tell rather in favour of 
this view may be briefly stated as follows. Ferns which have a 
continuous internal endodermis throughout the stem and no internal 
phloem are decidely rare. We have such a case, however, in the 
Gleicheniaceous genus Platyzoma (pp. 140—143), and the stele of this 
Fig. 90. Osmundites skidegatensis. Diagram of part of stele, leaf- 
gap with phloem continuous, l.g. 2., leaf-gap with xylem closed, l.t., leaf- 
trace not yet detached. After Kidston and Gwynne-Vaughan. 
plant also resembles that of the Osmundaceae in having no leaf-gaps 
putting the cortex in connexion with the pith, the leaf-traces 
departing from the outer surface of the xylem-cylinder somewhat as 
in the simpler Osmundaceous types. There is good reason for 
considering Platyzoma as an instance of reduction. In the Vittarias 
and Antrophyums, also, we apparently have a reduction-series 
culminating in the complete disappearance of internal phloem, while 
in some forms it remains only in the neighbourhood of the leaf-gaps 
(pp. 192—3). Scliizcea and the Ophioglossacese might also be cited 
