447 
Scott.—The Structure of Mesoxyloh multirame. 
secondary wood. Having reached the pith the strands turn down again and 
are cut almost transversely. They are converging somewhat, and between 
them, separated from each by a medullary ray, is the median strand, now 
a good deal narrower. The centripetal wood of the leaf-trace strands is 
well developed, and each is provided with a thick sheath of radially arranged 
elements, some of which are evidently tracheides. The median strand has 
no sheath and it is doubtful whether any of its xylem is centripetal, though 
in other cases it appears to be so. 
2765. The two strands are now flush with the inner edge of the wood 
and are about the width of a strand apart. The median strand is much 
reduced. 
2766. They are somewhat nearer together ; the median strand has dis¬ 
appeared, the two leaf-trace strands are separated by a ray with intercalated 
narrow radial bands of tracheides. The sheath has practically died out. 
2767. The strands are here close together, but still quite distinct, with 
their centripetal wood still well developed. 
2768. Little or no change. At this level the double trace still projects 
into a bay of the wood. 
2769. The double trace is prominent, as before, while the two centri¬ 
petal groups are still distinct, though reduced ; the centrifugal part of the 
xylem is no longer in two parts, but consists of six or seven radial bands, one 
of which is median. 
2770. The double trace is still slightly prominent with remains of the 
two centripetal groupsl The centrifugal xylem is scarcely marked off from 
the rest of the secondary wood. 
2771. The trace is no longer recognizable. 
It will be seen, on comparing the description of the course of a leaf-trace 
in the type-specimen, that there is no essential difference. Here also there 
is never an actual fusion of the two strands ; they merely become merged in 
the general zone of secondary wood ; the centripetal xylem-groups persist 
and remain distinct as long as the trace retains its individuality. The 
sheath, only seen where the twin bundles first reach the pith, is better nrnrked 
here than in the type. The leaf-traces are very numerous. In one trans¬ 
verse section (PI. XII, Fig. 15) 13 can be recognized in different parts of their 
course, and on following the series through the successive sections, many 
others make their appearance ; the total number found was 28, but possibly 
they may not quite all represent distinct orthostichies. However, the actual 
number was undoubtedly greater, as the sections are incomplete. As already 
mentioned, the phyllotaxis, while not exactly determinable, was evidently 
a complex spiral, with a most probable divergence of J|. 
The axillary steles are not well represented in the transverse sections; 
they are better developed in the part of the stem shown in the longitudinal 
sections, cut above the transverse series. The best of those in the trans- 
Gg 
