435 
1907-8.] Origin of adaxially-curved Leaf-trace in Filicales. 
there divide into two separate groups (fig. 6) — eventually, indeed, into three 
or more (fig. 7). Usually the protoxylem does not divide until the island 
has already become a bay, but it may do so before, while the island is 
still closed in by centripetal xylem. Referring in general to the ontogeny 
of the leaf -trace of the Ferns, it may be shown that any change in form is 
first of all initiated by the xylem strand and then followed by the outline 
of the trace as a whole. It is interesting to note that, while the changes 
described above are taking place in the xylem of the leaf-trace of Tham- 
nopteris, the outline of the whole trace still remains elliptic or oblong. It 
is not until the sclerotic outer cortex of the stem is reached that a slight 
depression appears on the adaxial side of the leaf -trace, which then rapidly 
becomes reniform in transverse section (fig. 7). Once the leaf -trace has 
actually entered the sclerotic cortex, the xylem strand becomes much 
narrower and more extended, and the protoxylem groups increase con- 
siderably in number. It also gradually becomes more and more curved, 
first taking the form of a low, rounded arch (fig. 7), the ends of which then 
slowly approach each other (fig. 8), until at last the characteristic horseshoe- 
shaped strand with incurved ends is produced (fig. 9). This curvature is 
followed by the outline of the leaf-trace as a whole — always, however, 
lagging a little behind that of the xylem. 
It -is, of course, to be understood that the changes described above are 
those that take place in the individual leaf -traces of the mature leaves as 
they pass through the cortex of the stem to the free petioles. Nevertheless 
it is . believed that these stages may be taken as indicating the changes 
undergone in the ontogeny and phylogeny of the leaf-trace, and that, 
therefore, they offer useful and reliable suggestions as to the origin and 
derivation of the adaxially-curved leaf-trace so representative of the 
Filicales. 
The essential points of this theory of the direct or normally oriented 
horseshoe leaf- trace are as follows. First of all, the xylem strand is derived 
from a primitively solid, more or less rounded mass with a central mesarch 
protoxylem. Secondly, this mass became concave on the adaxial side by 
the substitution of parenchyma for the centripetal elements of the xylem, 
the protoxylem thereby becoming truly endarch. This view is advanced as 
an alternative to the suggestion put forward by Mr Tansley,* who derives 
the directly oriented horseshoe from a Zygopterid type of leaf -trace by 
the complete reduction of the abaxial arms and the pinnae that they supplied. 
Without disputing the possibility that this method could have produced a 
* Tansley, A. G., “Lectures on the Evolution of the Filicinean Vascular System, 5 ’ New 
Phylologist , vol. vi., p. 64, 1907. 
