150 A. G. Tansley. 
of the insertion of a broad C-shaped trace upon a protostele is not 
difficult to understand. The current of water passing up the 
rhizome and partly deflected into the trace will take a peripheral 
course and the tracheal elements of the stele opposite the middle 
of the trace will tend to be depleted of water and eventually cease 
to be developed. Thus the tissue in the concavity of the trace will 
tend to be decurrent into the stele, and the “ pocket ” seen in 
Gleichenia will be developed. At the same time any broadening of 
the span of the leaf-trace arch will tend to necessitate an increase 
in the diameter of the stele in order to provide for its insertion. If 
the xylem of the stele remained solid it would increase as the 
square of the diameter of the enlarged stele, but this would be in 
excess of the additional demand, and would again lead to the 
replacement of the central tracheae of the stele by passive tissue, 
and still further to localise peripherally the xylem of the stele. If 
the central passive tissue is continued down through the internode 
and joins the decurrent strand of the node below we have a hollow 
stele developed 1 . Thus the evolution of the solenostele seems to 
depend primarily on basipetal evolution from the point of insertion 
of a C-shaped leaf-trace. 
Another factor may come into play in the development of a 
pith consisting of sclerenchyma. The insertion of a solid rod of 
resistent sclerenchyma in the node, one end of which supports the 
leaf-trace, while the other runs down the stele of the stem, must 
tend to strengthen mechanically the junction of the two, which 
will be a point of weakness in the continuity of the tracheal 
channels. Though this junction certainly seems to be quite 
adequately protected by the continuous external plate armour of 
cortical sclerenchyma in most of the existing types, the internal 
sclerenchyma will add to the rigidity of the whole system. 
The existence of the Lindsaya -type not only in the adult 
Lindsayeie, but also in the ontogeny of most of the more advanced 
Fern-stems that have been investigated, shows that the decurrency 
of internal phloem often precedes the decurrency of internal 
endodermis and sclerenchyma. It is easy to see that the evolution 
of a larger leaf-trace in order to supply a more complex frond will 
involve a demand for an increased conduction of formed food- 
material away from the frond pari passu with increased water 
conduction. Internal phloem thus developed in the concavity of 
the leaf-trace, at first by extension round the ends of the C, would 
1 The explanation here given is substantially the same as that 
suggested by Boodle (’01 A). 
