748 PROFESSOR W. 0. WILLIAMSON AND DR. D. H. SCOTT ON THE 
as we have seen, the number of perimedullary strands is either equal to that of the 
cortical leaf-traces or not greatly in excess of it. 
This we can explain in three ways : 
1. The leaf-traces on joining the cylinder may have branched, or 
2. They may have extended down through a very large number of internodes 
turning aside sufficiently to make room for each other, or 
3. It may be that some of the peripheral strands of the cylinder are cauline and 
not directly continucus with the leaf-trace bundles. It is not possible to decide 
between these suppositions ; the strands in question are not isolated like those of 
Lyginodendron, but form part of a continuous mass of wood. Hence their course 
could only be determined with certainty by tracing the actual protoxylem-elements 
through a series of sections, and that is more than we can expect to do, though in 
individual cases the protoxylem is clear enough. 
We must, therefore, be content to sum up our knowledge of the course of the 
bundles in Heterangium Grievit, as follows :— 
1. The leaves were spirally arranged, the phyllotaxis being $ in some of the larger, 
and 2 in some of the smaller stems. 
2. Each leaf received a single trace. The traces passed very gradually through the 
cortex, and joined the central cylinder at a distance below the node equal to from 6 to 
10 internodes. 
3. The traces can be followed for some distance downwards at the periphery of the 
central cylinder. 
3. The Primary Structure of the Stele and Leaf-trace Bundles. 
In most specimens the trachez occupy the larger part of the whole area of the 
trausverse section of the primary wood.* In some cases the stele is more paren- 
chymatous and the scattered trachez form a relatively small part of the whole tissue.t 
In no case, however, is there any sign of a definite central medulla. The paren- 
chymatous bands extend to the periphery of the primary wood, where they separate 
the more external xylem-strands from one another, and are continuous with the 
principal rays of the secondary wood (see fig. 24, c.p.). 
The conjunctive parenchyma is made up of short, thin-walled, square-ended cells 
(see fig. 25, ¢.p.). 
In the specimens of H. Grievw from Burntisland we found no other elements in 
this tissue; in the stems from Dulesgate, Lancashire,{ however, the conjunctive 
parenchyma is traversed hy rows of elongated cells with dense carbonaceous contents, 
similar to the structures which we have termed “secretory sacs ” in Lyginodendron. 
* Wituiamson, Part IV., Plate 28, fig. 30. 
+ ON. 1254. 
t Wixitamson, “ Organization,” Part XVII., p. 96. 
