692 Browne. — Contributions to our Knowledge of the 
/ 
of the annulus are difficult to observe in E. arvense. In Cone A of this 
species the meshes below the first and second traces of the lowest whorl of 
sporangiophores are closed, and a fresh mesh of the first order originates 
between them (cf. Text-fig. 1). In this restricted region the branching 
gives rise to the formation of a mesh alternating with that of the internode 
below, and the nodal appearance is most clearly marked in a longitudinal 
reconstruction, though, of course, the fresh meshes are not subtended by 
traces. In other cases a mesh is closed in the neighbourhood of the 
insertion of the annulus, but no fresh mesh originates higher up alternating 
with it ; this may, perhaps, be due, as in E. palustre , to the persistence of 
a band of xylem through the internode above the annulus. This seems to 
be the case with certain of the strands of Cone C of E. arvense. Here the 
first, the second, the third, the fourth traces, the two bundles composing the 
prematurely divided sixth trace, and the eleventh trace, are all inserted not 
vertically above the middle of the strands of the uppermost vegetative 
internode, but lie more or less above the point of junction of two strands. 
Thus the meshes above these traces do not alternate with those of the 
uppermost vegetative internode, but lie more or less above them ; on this 
assumption the meshes corresponding to the third and fourth traces have 
‘ decurred ’ a little below the traces to which they phylogenetically belong. 
If parenchymatous meshes had appeared in this band of xylem above the 
annulus and had been closed again at the insertion of the sporangiophores 
they would be about half the height of the meshes of the first order of the 
icone. In Cone A, where two meshes actually arise above the annulus and 
are closed before the insertion of the basal whorl of the cone, their height 
is about half that of the meshes of the first order found in the lower and 
less immature region of the cone. The absence of branching in the region 
of the annulus in Cone B may perhaps be due, at any rate in part, to this 
same relatively great development of xylem in the internode above the 
annulus. The xylem is much more developed in this cone relatively to its 
size than in Cone C, and at the level of the basal whorl all the strands are 
united into a ring of xylem. But the fusion of the xylem of the bundles 
forming a ring takes place at very different levels, and some of the strands 
fuse as far down as 0-9 mm. below the sporangiophores ; if meshes origi- 
nated just above this lowest fusion and were closed at the height of the 
sporangiophores they would be, like the meshes of the first order arising 
above the annulus of Cone A, about half the average height of the meshes 
of the first order of the cone of the specimen in which they occur. In 
Cone B, however, it was difficult to be sure if the meshes above the 
sporangiophores were superposed to those of the internode below, for the 
diameter of the stele narrowed abruptly at and above the insertion of the 
basal whorl of the cone, and this narrowing confused the relations of 
alternation and superposition of the meshes. It is quite possible, therefore, 
