Anatomy of the Cone and Fertile Stem of Equisetupi . 691 
closed at and fresh ones originate above the node. A branching of strands 
independent of the formation of a parenchymatous mesh above a trace that 
has departed is very rare. It is true that in cones with relatively little 
xylem, or in parts of cones with little xylem, a mesh may arise below and 
to one side of a trace ; but in such cases its form and position seem to show 
that it has, phylogenetically speaking, 5 decurred’ downwards below the trace 
to which it corresponds (cf. p. 673). True branching of strands, independent 
of the formation of meshes above traces, does occur, though rarely, both in 
cones and stems where the number of members in the next whorl is 
increased, and still more rarely owing to inexplicable irregularities. But 
even in cases in which an increase in the number of members causes branch- 
ing of the strands the area of disturbance of superposition of meshes to 
traces is narrowly restricted. Branching of some of the strands above the 
annulus is common, and strongly suggests that we are dealing with a reduced 
node ; in E. arvense , however, the number of sporangiophores in a whorl 
often considerably exceeds the number of leaves in the fertile stem, and the 
formation of fresh meshes above the insertion of the annulus might be 
regarded as the expression of the increase in the number of appendicular 
members. In this species, too, the number of strands that pass uninter- 
ruptedly through the annular node is relatively large. The transitional 
region was studied in Cones A, B, C, F, G, and H of this species. In 
Cone A four out of nine strands forked, giving rise to fresh meshes above 
the annulus. Below the annulus of Cone B there were seven strands, and 
all passed through the annular region without branching ; in Cone C one 
only of the eleven strands branched above the insertion of the annulus. 
Cone F was exceptional and will be considered later ; in Cone G three out 
of nine strands forked above the insertion of the annulus, while in Cone H 
one of eleven strands forked. The forking- observed in this last cone was 
o 
the more significant as there were only ten sporangiophores in the lowest 
whorl of the cone ; thus the branching of the strand was not correlated 
with an increase of appendicular members, and seems clearly to be a 
vestigial nodal character. In Cones A, B, and D of E. palustre , too, the 
number of members in the uppermost vegetative whorl was equal to the 
number in the lowest whorl of the cone ; yet in all these cases a certain 
number of meshes were closed, and others originated in the neighbourhood 
of the annulus. In the four cones of E. palustre examined, after the closure 
of certain of the parenchymatous meshes in the region of insertion of the 
annulus, a band of xylem which effected this closure persisted upwards 
through the ‘ internode 5 lying between the annulus and the sporangiophores ; 
in this region the absence of meshes was due, as it appears occasionally to 
be due in the cone (cf. p. 671), not to poor development of nodal wood, but 
to unusual development of internodal wood. 
Cases of closure of meshes of the internode below in the neighbourhood 
Zz 2 
