Anatomy of the Ophiogiossaceae. /. 223 
thickening its xylem is of course wholly primary from the beginning 
(Phots. 1 6, 17). 
No trace of centripetal xylem has been seen to the inside of the leaf- 
trace when it separates from the stele of the stem, 1 even when the latter 
shows remains of the central xylem (Phots. 16, 17). The protoxylem of 
the departing trace abuts directly on the medullary parenchyma. The 
layer of cells to the inside of the xylem-ring can often be distinguished as 
of different appearance from the other cells of the pith, and this layer of 
cells is always clearly recognizable on the adaxial face of the xylem of the 
departing leaf-trace. As the trace separates the next layer of cells may 
assume endodermal characters, and so complete the endodermis of the leaf- 
trace adaxially (Phot. 18). Whether a complete endodermis is present or 
not the adaxial parenchyma can be recognized as a part of the leaf-trace ; 
it usually increases by cell-division as the trace departs, so that the outline 
of the latter becomes oval. As the trace widens out on its passage through 
the cortex it assumes a rather flattened C-shape, the concavity being filled 
up or lined by the adaxial parenchyma. No phloem has been observed on 
the adaxial face of the leaf-trace. 
The most interesting variant of the usual structure of the leaf-trace as 
described above is the development of tracheides on the adaxial side of the 
protoxylem ; indications of these are found in most cases, and sometimes 
the adaxial xylem is well marked. The metaxylem of the trace then 
appears to curve round at the two ends, while more scattered tracheides are 
developed across the adaxial face of the trace joining the two ends 
(Phots. 18, 19). The adaxial tracheides are seen in longitudinal section in 
Phot. 20. The adaxial xylem is most marked when the trace is passing 
through the cortex of the stem, but occasional tracheides have been met 
with after the trace has entered the petiole and divided. A common state 
of affairs is for only a slight adaxial extension of xylem to take place at the 
two ends of the trace. This may probably be related to the tendency of 
the xylem of the trace to form a hook, as is shown in Phots. 21 and 27. 
Such leaf-traces might be described as mesarch, but it is questionable 
whether this is a sufficient or right interpretation. It has been pointed out 
that the trace is never truly mesarch at its departure, the central xylem, if 
present, disappearing before the trace departs. It is, of course, possible to 
regard the adaxial xylem in the trace as a reappearance of the central or 
centripetal xylem, in which case the trace would be truly mesarch. But it 
seems to be better regarded as due to a more or less complete adaxial 
extension or completion of the centrifugal xylem of the trace ; this produces 
a trace the xylem of which has the form of a ring enclosing more or less 
parenchyma. It should be mentioned that, even when the endodermis and 
xylem are thus completed, no corresponding completion of the phloem has 
1 Except in the case of one branching plant described below (Text-fig. 14). 
R 2 
