OP THE FOSSIL PLAHT-S OF THE COAL-MEASURES. 
383 
In the innermost layer just described we find large longitudinally disposed vascular 
bundles, the presence of which, under various conditions, constitutes one of the charac- 
teristics of the genus Dictyoxylon. These bundles sometimes, though rarely, occur 
singly ; but even when they do so we can generally trace a radial line, indicating a 
division of the bundle into two lateral halves. An example of this latter condition is 
seen in Plate XXIII. fig. 7, z. More ordinarily we find two such bundles (Plate XXY. 
fig. 17, z), evidently resulting from the vertical division of a cluster, like that just 
referred to, into two nearly equal parts. The transverse section of each bundle is now 
somewhat of a rounded triangle in form, the semiovoid form being due to the same 
^ vertical division of an oval double bundle. 
Such undivided bundles as that of Plate XXIII. fig. 7 have a longer diameter of ‘033. 
When the two halves become separated, as in Plate XXV. fig. 17, by a decided and 
sometimes considerable interval, each half attains to about the same dimensions as the 
undivided whole, making it clear that as the division progresses the component vessels 
increase in size. The diameters of these vessels vary from ’003 to ’0015. On examining 
the bundles in a longitudinal section (Plate XXIV. fig. 11, z), we observe that they 
preserve their parallelism with the exterior of the ligneous zone, and do not seem to 
take their rise in, or diverge from, the latter structure, as is the case with the foliar 
bundles of the Lepidoclendroid plants, with which, at the first glance, we should be 
tempted to compare them. Their component vessels are chiefly reticulated ones, like 
those of the woody zone ; but along with these are other and usually smaller ones, 
which are distinctly of the barred type. Such barred vessels are, as I have already 
shown, entirely absent from the true ligneous zone. 
In the great majority of cases I have counted four of these double bundles in each 
transverse section of the entire stem. In Plate XXII. fig. 1, I only discover three pairs, 
whilst in Plate XXII. fig. 4, 1 have reason to believe that there were as many as five. The 
positions of these double bundles bear a remarkable relation to those of the separate masses 
of the original medullary cylinder (Plate XXII. figs. 2, 3, & 4, & Plate XXIII. figs. G & 7, c ), 
which remain attached to the medullary surface of the ligneous zone. Where the relative 
positions of the organs have not been disturbed by mechanical pressure, as is often the 
case, these cortical bundles are disposed externally to the woody zone at points inter- 
mediate between those of the medullary clusters of vessels. I have seen no exception 
to this rule. Of course, this alternation is less obvious in young stems, in which, like 
Plate XXII. fig. 2, the medullary cylinder is not yet broken up into distinctly separated 
portions, than in older ones, though even here, as seen in Plate XXIII. fig. 7, the 
evidence of this relationship is not wanting ; but where, as in Plate XXII. figs. 1 & 4, 
cylinder. Whilst referring to the necessity for incessant vigilance against this source of error, I may express 
my absolute conviction that the vascular bundles which Mr. Binxey has figured as belonging to the medullary 
tissues of Halonia (“ Observations on the Structure of Fossil Plants, &c.” (Palseontograidiical Society, 1871), 
■plate xvi. figs. 2 & 5, and plate xvii. fig. 5) are merely examples of these intrusive rootlets. 
