316 Holden. — On the Anatomy of 
or oblique walls (Fig. 4). Both D . Pedroi and D. lafonense have somewhat 
similar secretory elements, but they seem more in the nature of canals, the 
surrounding parenchyma cells being converted into a pseudo-sheath. These 
in question resemble rather the isolated cells of Lyginopteris . 
Inserted between the pith proper and the vascular stele, there is a 
jacket of cells of a different nature. It varies in width, but generally speak- 
ing is best developed at the nodes. The smaller diameter of these elements 
serves to differentiate them in cross-sections (Figs. 2 and 7, PL XVIII), while 
longitudinal sections reveal their characteristically reticulate, tracheoidal 
markings (Figs. 6 and 17, PL XIX). They seem, in fact, to be identical 
with the transfusion tissue found in the leaves of Conifers, 1 Casuarinales, 2 
&c. Instances of their presence in stems are less common, but Rothert 3 re- 
ports them in Cephalotaxus and Thompson 4 in Ephedra . In all these cases, 
however, they occur as isolated cells, or small groups intermixed with 
parenchyma, rather than as a solid sheath, while in Megaloxylon 5 they are 
not limited to the periphery, but extend throughout the pith. For strictly 
comparable structures we must go to Antarcticoxylon 6 and to Mesoxylon 
Lomaxi and M. platy podium? 
Leaf -traces. 
As regards leaf-traces, the absence of serial sections through the node 
renders it impossible to ascertain the conditions with exactness. It seems 
fairly certain, however, that the internodes were variable in length, a long 
one being followed by several very short ones. The result must have 
been that the leaves appeared superficially to be in whorls, but tangential 
sections prove that each was separated from the next by an appreciable 
distance. The course of the traces through the wood is almost horizontal 
until they approach the pith, when they bend steeply downward. The 
most striking feature is that they are always in pairs. A double trace 
is characteristic of practically all the Pteridosperms and Cordaitales, 
with significant variations. The most primitive condition is probably that 
of Lyginopteris , Heterangium^ Medidlosa , and Calamopitys , 7 where the 
trace leaves the stele as a single strand and bifurcates either once or 
repeatedly in the pericycle and cortex. The next stage is represented 
by Poroxylon , 8 where the division takes place some time before the trace 
leaves the margin of the pith, and the two strands pursue their outward 
course through the tissues of the secondary wood in close juxtaposition. 
Mesoxylon represents the next stage, for here not only are the two strands 
distinct at the margin of the pith but, as they penetrate the woody stele and 
pass out, they diverge more and more. Finally we come to our Indian 
1 Jeffrey (1905). 2 Boodle and Worsdell (1894). 3 Rothert (1899). 
4 Thompson (1912). 5 Seward (1899). 6 Seward (1914). 
7 Scott (1912). 8 Bertrand (1889). 
