1084 
Thompson . — The Anatomy and 
such element separated by at least two pith cells from the xylem. By 
following a series of sections one finds that they occur in strands attached 
to the protoxylem below, but diverging centripetally above. The lower- 
most elements are most like tracheides, but the upper ones rapidly assume 
the characters of transfusion cells. It is probable that they represent the 
original centripetal wood which was present universally in the ancestral 
types, and has been lost from the internode as in so many other plants, but 
in this case retained in a modified condition at the node. 
At the edges of the primary bundles of the internode one often finds in 
E. monostachya and E. vulgaris typical transfusion tracheides like those of 
the leaf. They are adjacent to the pith, and often reach across from one 
bundle to the next. In Fig. 9 two primary bundles may be seen at the 
extreme right and left, and between them three or four of the scattered 
transfusion tracheides with their characteristic bordered pits and thickening 
bars. Their presence is probably due to the fact that the cortex of the 
young plant functions as a leaf. 
Having described the origin of the leaf-traces in connexion with the 
course of the primary bundles, it may be well at this point to take up 
its structure. In all cases the trace is very small, consisting of but a few 
elements. Its tracheides are nearly all spiral or scalariform, although occa- 
sionally one may find a pitted element. In all cases the arrangement 
is endarch, i. e. the protoxylem elements are on the adaxial side of the 
bundle, and the metaxylem is all centrifugal. The endarch character may 
be distinguished in Fig. 7, in which the protoxylem elements of the trace 
are uppermost. In no case was centripetal xylem found. The structure is 
then identical with that of the trace of the Angiosperms and Conifers. 
It is vitally different from that of the Cycads and older Gymnosperms. In 
the Cycads, as is well known from the work of Scott and the earlier studies 
of Mettenius, there is always a large amount of centripetal wood which 
is considered to be retained from a primitive condition in which it occurred 
in the stem as well as in the leaf. At the same time the centrifugal primary 
wood is very small in amount and diminishes in the petiole. In Ephedra 
the conditions are exactly reversed ; the centripetal wood has been com- 
pletely lost and the centrifugal wood developed. In regard to the structure 
of the leaf-trace, therefore, Ephedra is as remote as possible from the 
Cycads, and is in agreement with the Conifers and Dicotyledones. In view 
of the general agreement as to the importance of the leaf-trace as a seat of 
ancestral characters, this is a point which should be emphasized in all 
phylogenetic studies. 
Secondary Wood. 
It has long been known that the secondary wood of Ephedra consists 
of tracheides, vessels, wood parenchyma, and rays. But an intimate study 
of each of these constituents is necessary, especially in the light of recent 
