44-6 Scott .—- The Structure of Mesoxylon multirame. 
actual fusion. Axillary steles are present in connexion with some of the 
leaf-traces. 
The points in which this specimen throws further light on the structure 
of the stem will now be indicated. 
While the middle part of the pith has perished, only the outer ends of 
the diaphragms remaining, the persistent external zone is remarkably well 
preserved. The layers next the wood have somewhat smaller cells than the 
rest, usually with dark contents, perhaps indicating the presence of starch 
reserves during life (PL XII, Figs. 14, 15). The persistent pith cells gener¬ 
ally are isodiametric in transverse section, and appear somewhat flattened 
(broader than long) when cut longitudinally. There is little elongation 
even of the peripheral cells. In these respects the specimen differs some¬ 
what from the type (see above, p. 439). The inner layers show evident 
tangential divisions, suggesting that a medullary periderm was being formed 
as a barrier towards the central cavity, which, apart from the diaphragms, 
must have existed during life. The same is the case where the continuity 
of the persistent zone is interrupted by a radial split. The frequent trans¬ 
verse splits in the persistent zone, however, show no sign of periderm forma¬ 
tion, and were evidently due merely to post-mortem shrinkage, which may 
account in part for the flattened form of the cells. 
The details of the course of the leaf traces are shown even better than 
in the type-specimen. A typical leaf-trace, with its axillary stele, shows the 
following changes, followed from above downwards. 
The trace chosen first appears in the second section, 2761. Here two 
bundles (one of which is double) are seen somewhat obscurely in oblique 
section entering the periderm. In the wood opposite this point a double, 
nearly horizontal, strand is conspicuous, the two parts of the strand 0*4 to 
0*45 mm. apart. From evidence adduced below, there is no doubt that this 
represents the axillary stele of the trace, which is known to have often 
divided into two in its passage through the wood. 
2762. Here the two bundles of the trace, about 1*5 mm. apart, are 
seen clearly in transverse section in the periderm ; one of them is still 
evidently double. The axillary stele, as such, has disappeared, having 
already fused with the wood, at this level; but its place is marked by 
a conspicuous bay in the inner edge of the wood, into which a median strand 
projects. This is always the indication of an axillary stele having just 
passed in. 
2763. The two bundles of the trace have reached the phloem, and are 
about 1 mm. apart; at the inner edge of the wood the bay, with the median 
strand, is very conspicuous. 
2764. The trace has passed rapidly inwards and the two strands have 
now traversed the wood and reached the bay. This quick passage is of 
course due to the almost horizontal course taken by the trace through the 
