of Vascular Strands in Angiopteris evecia . 377 
lattice work differing only in the size of the mesh work from 
that of a typical * polystehc ’ fern. 
Meanwhile the leaf-traces, as has been said, increase in 
size, and at the same time the dichotomy above referred to 
extends farther and farther towards their bases till it becomes 
apparent almost as soon as they leave the central strand, or 
as we prefer to term it, siphonostele. The roots frequently 
originate directly from the siphonostele, though oftenest not 
from the margin of a gap, but rather from a broader part of 
the vascular tissue ; they may, however, also spring directly 
from a leaf-trace, and in older stems they become gradually 
more intimately related to the latter* though a few may still 
spring from the axile strand itself. The latter series pursue 
a sinuous course through the parenchymatous ground-tissue 
of the stem, finally passing out through a foliar gap into the 
outer cortex. 
The skeletal tissue now rapidly assumes a more complex 
character, arid begins to exhibit the features appertaining to 
the adult form, after which the increasing thickness of the 
stem only produces further complication on the lines thus 
early inaugurated. 
The traces of all but the earlier formed leaves become 
double from their point of exit from the siphonostele, and 
above each pair there is a huge foliar gap. The commissural 
strands which traverse the axile parenchyma of the siphono- 
stele become more obvious, though they still are very slender 
as compared with the strands of which the cylinder still is 
made up after the exit of the traces. The latter become 
irregularly disposed and appear at this stage to arise from 
the edges of the gap, an appearance due to the deep dicho- 
tomy. They also have begun to divide up farther before 
entering the petiole. 
Finally, the siphonostele opens out to a considerable width 
(Fig. 6), whilst the axile commissures assume an ever-growing 
importance forming a sort of sympodial column. In fact, 
these large commissural strands may easily be misinterpreted 
as representing the original stele. It is this strand which 
