682 Browne . — Contributions to our Knowledge of the 
alternation there are three forms : (i) the xylem-strand of the next inter- 
node may be formed by the fusion of a branch of a forking strand with 
a (whole) strand of the internode below (Text-fig. 7, c) ; or (2) by the fusion 
of two whole strands (Text-fig. 7, d) ; or (3) a strand may fork above the 
departure of a median trace and one (Text-fig. 7,e) or more, rarely both 
(Text-fig. 7, F), branches may constitute, independent of any fusions, trace- 
bearing strands of the internode above. The first form of irregular alterna- 
tion is by far the commonest in E. limosum ; the second leads to reduction 
in the number of strands, and is therefore more characteristic of the upper 
part of the cone ; the third naturally leads to an increase in the number of 
strands in the next whorl unless it is counteracted by the fusion of two whole 
strands of the internode below. Such irregularities in alternation, when not 
caused by a change in the number of members of successive whorls, are due 
to failure to close some of the meshes at the nodes, and therefore to poor 
development of the xylem in this region ; a further reduction of xylem leads, 
as already pointed out, to the formation of strands pursuing an isolated 
course through the nodes, and therefore to the superposition of traces. 
A case occurred in Cone C of E. limosum in which a xylem-strand 
appeared to branch without any relation to the departure of a trace. 
A narrow strand, after giving off a trace, comes temporarily into connexion 
with one of its neighbours, about three-quarters of the way up the internode 
(and thus closes a parenchymatous mesh) ; but this connexion only extends 
for a short distance, and less than half-way up the next internode the strand 
again detaches itself. 
Another exceptional, apparently a unique, irregularity occurs in the 
lowest whorl of Cone C. Here all the xylem-elements above a xylem-trace 
do not die out ; but some of them pursue an oblique course and fuse with 
a neighbouring strand, thus closing a mesh of the internode below and 
initiating one in the internode above. 
The lowest whorl in Cones A and C shows further anomalies ; in both, 
while certain strands pursued an isolated course through the node and gave 
off no traces, pairs of strands united to one another at the node gave off 
three traces. 
Apex of the Cone. 
The structure of the apex of the cone differs in the different species. 
Externally, the terminal sporangiophore of E. arvense is larger than the 
others, and the outer (in this case upper) surface of the shield of the 
sporangiophore, instead of being flat or slightly depressed in the middle, is 
prolonged into a drip point ; this relatively massive terminal structure 
never has, as the terminal sporangiophores of E. palustre and E. limosum 
have, a single more or less circular trace. Cone A was so very young that 
the terminal internode had not elongated at all, and the apical phenomena 
