66 Moslen . — 7Yz<? Relation of Root to Stem in Catamites. 
The longitudinal sections are cut parallel to the longer diameter of 
the transverse section of the flattened stem, and the series passes completely 
through the xylem of one side, then across the pith and the wood on the 
other side, and finally into the matrix in which the stem is embedded. 
The sections pass through one node, and at least five roots can be seen 
springing from this node. One of the roots branches when it is free from 
the wood of the main axis. 
There is very considerable difficulty in distinguishing the stem-branches 
from the root-branches in sections which are taken close to the insertion 
of the lateral members on the primary wood of the parent axis, though 
it is usually easy enough when they are cut farther out. 
Good figures of stem-branches are given in Williamson and Scott’s 
Memoir 1 , and they show, while still embedded in the xylem of the main 
axis, a relatively wide pith surrounded by a ring of vascular bundles each 
provided with a small canal (carinal canal) marking the positions of the 
protoxylem groups. The presence of the canals renders the identification 
of such sections as stems and not roots easy. The difficulty increases with 
the nearness of the branches or roots to their seat of origin, especially since, 
as Williamson and Scott have shown, the characteristic inter-nodal or carinal 
canals, which mark the position of the disorganized protoxylem groups of 
the Calamitean stem, are not present at the actual base of the branch 2 . 
Given then an isolated section passing transversely or longitudinally 
through the base of a stem-branch or an adventitious root, it is usually 
practically impossible to determine which of the two it is. A number 
of serial sections cut from the same specimen may, however, enable one to 
follow the stem or root outwards and so to determine its character. With 
regard to the position of the stem-branches, Williamson and Scott have 
shown that they are placed immediately above the node (i.e. the level of 
the leaf-traces), and usually between two of the bundles coming in from the 
leaves. 
Plate I, Fig. i (slide 551 S) is from a tangential longitudinal section 
of the stem showing two roots in approximately transverse section. The 
main axis is evidently a stem as it shows clearly the usual bifurcation 
of the bundles at the node, leaf-traces (/. t.) passing through the nodal wood, 
and Williamson’s ‘ infra-nodal organs ’ (i. n. o.). As the latter always occur 
below the node 3 they enable one to orientate the sections correctly, i. e. to 
distinguish the upper and lower ends. The section is evidently cut near 
to the pith of the stem, and passes through the inner portion of the wood : 
1 Further Observations, &c., Part I, Calamites, Calamostachys, and Sphenophyllum, Phil. 
Trans., vol. clxxxv, B. 1895, PI. LXXII, Figs. 5 and 6. 
2 Loc. cit., p. 891. 
3 Small enlargements at the lower ends of the medullary rays (i. e. above the node) often occur, 
but these are usually easily distinguished from the infra-nodal organs. 
