440 Scott .— The Structure of Mesoxylon multirame . 
adaxial elements are oblique and in the act of joining those of the wood of 
the main stem. The bundles of the subtending leaf-trace are seen in the 
cortex, outside the periderm ; not all are preserved (five are visible), and it 
is possible that some fusion among the original eight had already occurred. 
A somewhat similar stage, but rather lower down, is shown, from another 
trace, in the photograph, PL XI, Fig. 3. In the latter case the leaf-trace 
bundles have already united so far as to form two symmetrical pairs. In 
both instances the lateral connexion of the wood of the axillary stele with 
the stem-wood on either side is well shown. 
In the next lower section (2337) the trace, No. 13, has passed within 
the periderm, and is entering the phloem. Its strands have now united into 
the pair of bundles characteristic of the trace in the lower part of its course 
(cf. PI. XI, Fig. 4, from another trace). The stem-wood behind the leaf- 
trace is still disturbed, owing to its junction with that of the axillary stele. 
A pith-bay intrudes into the wood at this level, and a definite median strand 
of xylem lies directly below the attachment of the axillary stele. 
In the next section below (2338) the double trace is just entering the 
secondary wood, the twin bundles passing in on either side of the median 
strand, with which they are connected by oblique elements. 
One section lower (2339) the twin bundles, converging slightly, have 
almost reached the pith. There is a space between them, almost equal to 
the width of a bundle, which is still occupied by the median strand. Almost 
the same condition, from another trace, is shown from a photograph in 
PI. XI, Fig. 5. The bundles at this level have well-developed centripetal 
xylem ; there is no definite sheath, though sometimes a few short tracheides 
are found in the tissue on the inner side of the bundles. 
At the next lower level (2340) the double trace has definitely taken up 
its position on the inner edge of the wood, bordering on the pith, into which 
the two strands project somewhat. They are nearer together than before, 
and the median strand between them is no longer well marked ; it stops 
short of the pith. The bundles of the trace retain their centripetal wood 
with little if any diminution. A similar stage is shown in PI. XIII, Fig. 19. 
One section below again (2342) the trace still forms a prominent pair; 
the twin bundles are quite distinct but approximated ; near the pith 
a biseriate ray separates them. The median strand has split up, and its 
tracheal rows on either side have joined the wood of the twin strands, from 
which they can no longer be distinguished. The position of the protoxylem 
is difficult to make out; but the centripetal wood is evidently much reduced, 
nearly the whole xylem belonging to the radial centrifugal series. 
The next section (2341) shows an interesting change. The two bundles 
of the trace are no longer distinct; the trace is here represented by a single 
broad xylem-mass projecting into the pith. The mass is made up of seven 
radial bands of tracheides, separated by medullary rays. There is no longer 
