of Vascular Strands in Angiopteris evecta. 387 
acid, and although in Angiopteris it might escape detection 
if only superficially looked for, we are at a complete loss to 
account for these statements, constantly repeated, as to its 
absence in the other two genera. K aulfussia indeed is even 
a somewhat favourable object on which to demonstrate its 
occurrence (Fig. 16), as it can readily be rendered visible by 
careful staining as well as by the use of sulphuric acid. The 
existence of an endodermis which is thus proved to be present 
in the Marattiaceae is a matter of some interest, as its 
supposed absence has been looked upon as an important 
differentiating character between this group and the Ophio- 
glossaceae. But after what has been said it is clear that no 
such line of argument can any longer be sustained. 
As has now been pointed out, the vascular tissue early loses 
its primitive simplicity : not only does the central strand ulti- 
mately become dialystelic (siphonostelic), but the leaf-trace 
strands become more complex, first by becoming forked, from 
the point of their origin, and in still more advanced stems 
by further forming an increasingly elaborate network of 
anastomoses by which the series of zones distinguished by 
Mettenius are produced. Meanwhile the commissural strands 
previously alluded to are formed across the central (pith) 
parenchyma that is still enclosed within the original frac~ 
tionating siphonostelic cylinder. These commissures at their 
first appearance are probably secondary in origin but they 
rapidly acquire a considerable degree of importance in relation 
to the peripheral vascular cylinder. The earliest formed com- 
missures may consist chiefly or even entirely of phloem, 
starting as a local hypertrophy of this tissue on the inner 
surface of the xylem. They become detached from it higher 
up ; and cross over to the opposite side of the siphonostele. 
Here they become reinforced by xylem elements, and a leaf- 
trace is commonly given off at the place where they join up 
with the tissues of the cylinder. It is, however, not very 
unusual to find some of these rudimentary commissural strands 
ending blindly in the central parenchyma, a fact which sup- 
ports the view as to their secondary origin in such cases. As 
