OF THE FOSSIL PLANTS OF THE COAL-MEASUEES. 
391 
a true exogenous ligneous zone. It appears to me that corresponding phenomena reappear 
in the Dictyoxylons, hut with some characteristic differences. In the Lepidodendroid 
plants each vasculo-medullary axis expanded into a regular cylinder, whose walls were 
of almost uniform thickness throughout its entire circumference, and which permanently 
retained its integrity throughout the life of the plant. In Dictyoxylon the solid vascular 
axes of Plate XXV. fig. 16 & Plate XXVI. fig. 23 expanded in like manner, and also had 
a cellular medulla developed within each cylinder; but the thickness of the latter was 
not only unequal in its various parts, but, at a very early period, it broke up into several 
detached and irregular masses, which no longer effected a complete separation of the pith 
from the exogenous layers throughout the entire circumference of the former organ. In 
the intervals between these detached vascular masses the cells of the medulla, and those of 
the medullary rays, become continuous with one another. In the first instance each 
.radiating exogenous lamina had its inner margin in direct contact with the medullary 
vessels ; this is indicated clearly in such specimens as that represented in Plate XXII. 
fig. 3. But though the disruption of the cylinder ultimately limited this direct cohesion 
of the exogenous with the medullary vessels to certain well-marked clusters of the former, 
evidence of similar but earlier relations of the entire ligneous zone is retained in the 
curvilinear forms of many of the woody wedges, whose inner margins are now in direct 
contact with the medulla ; as in fig. 3, their several medullary extremities still bend 
towards the nearest cluster of medullary vessels with which they were, at an earlier period 
of their growth, in direct contact. 
I have already called attention (in page 383) to the relations which these detached 
medullary bundles sustained to those lodged in the inner bark, and pointed out the way 
in which the latter helped to equalize the resisting power of the compound vascular 
cylinder, by strengthening those parts of the woody zone which were weakened by the 
breaking up of the medullary cylinder. This idea receives confirmation from the fact 
that, so long as the central axis is occupied by an undivided vascular bundle, these cortical 
buttresses are invariably wanting. In no one of the lateral diverticula which I have just 
described, and in which the solid axis invariably existed, did I find a trace of these cortical 
bundles : all the facts indicate that they were growths developed in the bark at a later 
date ; at the same time such specimens as the original of Plate XXII. fig. 2 showed that 
they became fully grown before the disruption of the medullary cylinder had proceeded 
to any considerable extent. 
The large size of the medullary rays, and the simple and conspicuous organization of 
the woody wedges in Dictyoxylon, enable us easily to compare its growth with that of 
living Dicotyledonous stems. As in the latter instances, we find that, as the wedges 
extended themselves in the peripheral direction, they became enlarged transversely by 
the addition of new laminse. This was sometimes the result of two meristem-cells occu- 
pying the place previously filled by one, the consequences of which are seen in the com 
version of the single lamina into a double one, as in Plate XXII. fig. 3, e. In other 
instances new and very small laminse were either intercalated into the substance or added 
