ic8 Sinnott.—Foliar Gaps in the Osmundaceae. 
strand is the appearance, usually near the outer margin of one of the xylem 
bundles, of a group of small protoxylem elements. Just inside this, a cluster 
of parenchyma-cells soon arises, which gradually increases in size towards 
the centre of the stem until it unites with the parenchyma of the pith, 
leaving the xylem-bundle in the shape of a horseshoe, with the protoxylem 
group near the middle of its inner face. The outer part, or curve, of the 
horseshoe now separates itself from the rest of the bundle and departs 
(with the phloem, which now surrounds it) as the leaf-trace, leaving the 
original bundle divided into two by a e ray * of parenchyma. This ray is 
a foliar gap in the xylem caused by the departure of a leaf-trace. The two 
bundles eventually unite, closing up the gap thus formed. 
This mode of separation of the leaf-trace from the cylinder was found 
to be by far the most common one in the group. Certain exceptions to it, 
however, were noted. In the stem of O. cinnamomea numerous instances 
were observed in which the cluster of parenchyma, after its appearance in 
the xylem-strand, instead of becoming continuous with the pith, broke first 
through the outer part of the xylem, on either side of the departing leaf- 
trace, and thus separated the trace from the original bundle. A transverse 
section cut through this region seems to show clearly a leaf-trace leaving 
the cylinder without causing a break in the continuity of the xylem. 
Continuous serial sections were cut through such places as this, however, 
and it was found, in every case where a complete series could be obtained, 
that the xylem-bundle was soon much constricted and finally broken into 
two by a ‘ ray ’ of parenchyma opposite the outgoing trace. This ‘ ray ’ is 
often only a few cells wide, but is nevertheless a true foliar gap, differing 
from the others only in its relatively late appearance. It soon closes up, as 
do the other gaps. PI. XI, Fig. i, shows a leaf-trace departing from the 
cylinder of O. cinnamomea^ apparently without making a break in the xylem. 
Fig. 2 shows the same trace at a slightly higher level. The gap opposite 
to it is now plainly discernible. 
O . regalis shows much the same condition of affairs. With most of the 
traces the gap in the cylinder is formed when the foliar strand separates 
from the stem-bundle, but in a number of instances observed the paren¬ 
chyma of the pericycle and of the pith came into communication only after 
the trace had progressed a considerable distance into the cortex. Fig. 3 
exhibits a trace just departing from the cylinder. In Fig. 4 the very narrow 
gap is shown which eventually appears in the xylem opposite this bundle. 
In O. Clciyto 7 iiana these cases of apparent absence of gaps, though 
occurring, are not as frequent, and almost all the traces go off in the usual 
manner. 
The mature stems of three species of the genus Todea were examined— 
those of T. bar bar a, T. superb a, and T. hymenophylloides. In the first of 
these, well-marked gaps were found in every case, and no temporary 
