678 Browne . — Contributions to our Knowledge of the 
The meshes originating below the cone and persisting into it will be 
considered in discussing the region transitional from fertile stem to cone. 
It has already been shown that where a strand pursues an isolated 
course through a node, the absence of a mesh above a median trace —or, in 
other words, the failure of a strand to branch above a median trace — is an 
effect of poor development of xylem. This is confirmed by the distribution 
of this type of median trace. In Cone A of E. limosum, the reduction in 
the number of members in a whorl begins above the seventh node ; in the 
seven lowest whorls, 31 median traces out of 40 subtended parenchymatous 
meshes. Similarly, the seven lowest whorls of Cone B (in which there was 
no definite reduction, though the number of the traces varied slightly in diffe- 
rent whorls) showed 27 median traces, of which 21 subtended meshes. The five 
lowest whorls of Cone C, which showed no signs of reduction in the number 
of members of a whorl, had 34 traces, of which 26 subtended parenchymatous 
meshes. But when we consider the regions of the cone in which reduction 
of the number of appendages has set in, we find very few parenchymatous 
meshes above median traces. This region was not available in Cone C, and 
in Cone B the region of reduction showed no median traces of either sort ; 
but in Cone A there were in the four uppermost whorls only three median 
traces subtending meshes, while above six such traces the strand pursued its 
course upwards without branching. As already pointed out, Cone B has 
relatively to its size much less xylem than Cone A ; in its upper part 
a large number of traces depart from strands so narrow that the traces are 
attached to the whole width of the strand. Such traces abutting on both 
edges of the strand cannot be called median ; but they clearly represent 
a further stage of reduction of the axial xylem to which a median trace not 
subtending a mesh is attached ; in the upper part of Cone B they replace 
the median trace. They are not, however, confined to the upper part of the 
cone, and an example may be seen in the eighth trace of the second whorl. 
No examples of such narrow trace-bearing strands were found in the other 
species or in Cone A of E. limosum. Some parts of the latter had rather 
more internodal xylem than others; this may cause the closure of the 
meshes a little way below the node ; or the greater development of the 
xylem may take place above the node ; for instance, the third and fourth 
traces of the fourth whorl of this cone depart from the edges of a band 
which persists upwards through the whole of the next internode ; if these 
two traces were median instead of lateral, the absence of a mesh above them 
would be due, not to the xylem being poorly developed, but to its being 
well developed. I have not, however, observed such a case in this species. 
Occasionally in E. limosum , isolated strands pass through a node with- 
out giving off a trace (e. g. two of the strands in the lowest whorl of Cone A, 
two of the strands in the lowest whorl of Cone B, two of the strands of the 
third whorl, and a strand of the fourth whorl of the same cone). In such 
