488 
Gwynne- Vaughan. — Observations on the 
The ancestral fern leaf is regarded as having a single row of branches 
on each side of the rachis. It follows, therefore, that the double row of 
branches on each side of the petiole possessed by some Zygopterideae is 
held to be a special development evolved within this affinity and one 
probably related to the assumption of an erect habit of growth by the leaf 
and the consequent arrangement of the lateral appendages in a more or less 
radial manner to avoid overshadowing. 
We may now consider the way in which the lateral branches of this 
primitive leaf may be assumed to have received their vascular supply from 
the rachis. It seems probable that, when about to give rise to a branch- 
trace, one of the lateral protoxylems would elongate in the direction of the 
long axis of the ellipse and then divide. A small mass of xylem would 
then protrude from the lateral margin of the trace, enclosing the outer 
protoxylem, and this would eventually be constricted off as the branch-trace. 
In the Zygopterid affinity this conjecture receives a considerable amount of 
support, because in the petiole of Asterochlaena^ which is on all grounds 
still fairly close to the common ancestor, a process very near to this is 
known to take place, the only complication being that the lateral islands 
of parenchyma are already present. 
It was to see whether the C-shaped leaf-trace still retained any features 
which might indicate the primitive method of branching that the following 
investigation was made. 
The manner in which the vascular supply of a branch is given off from 
its mother-axis in the Osmundaceous leaf varies in the same petiole accord- 
ing to the order and position of the branch in question* The simplest 
methods are to be found where the smaller veins depart from the midrib of 
the lamina or where small secondary pinnae arise on the rachis of a primary 
branch. In these cases the xylem strand of the trace is too small and thin 
to exhibit the features of special interest present in the branching of the 
stouter but still quite small traces to which our attention will be mainly 
directed. In the very simple branchings referred to above, the xylem 
strand of the trace may be reduced to an almost straight transverse band, as 
in Osnuinda Claytoniana (PI. XIII, Photo i), or it may still form a well- 
curved crescent, as in Todea barbara. In either case the endarch protoxylem 
on the side of the branching divides into two (Photo i) and the xylem strand 
constricts between the two protoxylems so as to nip off the outer of them 
for the trace of the branch (Photo 2). 
In somewhat larger though still quite small branchings, when the 
outline of the trace varies from elliptic to more or less reniform in transverse 
section and the xylem strand is a fairly stout and well-curved C with two to 
four groups of protoxylem, the procedure is most often as follows. As the 
1 Paul Bertrand: Structure des stipes C\' Asterochlaena laxa, Stenzel. Mem. de la Soeiete 
Geologique du Nord, t. vii, p. 1. Lille, 1911. 
