178 W.. T GORDON 
larger, and the protoxylem groups of the pinna-traces, vertically above the pair re- 
presented by the pinna-trace-bar we are considering, are quite distinct on the inner 
margin of the island. A somewhat similar example may be seen at the top of the trace 
in Pl. Il. fig. 20. There are at this stage four protoxylem groups arranged round the 
island. The groups at the ends of the island then spread out on the outer margin and 
afterwards divide into two. The stage just before the division is shown in the last- 
mentioned figure. 
Passing to the top of the trace in Pl. II. fig. 22, it will be noticed that the pinna- 
trace-bar has four protoxylem groups peculiar to itself. These have been derived as 
indicated above from the two original groups. There are thus six groups round the 
island, which has now attained a larger size. Two closely similar examples are figured 
in Pl. II. fig. 21 and fig. 18, at the top of the trace in each case. In figs. 17 and 19 the 
bar has become detached from the petiole-trace and gives some indication that it will 
divide into two equal parts. Pl. III. fig. 37 represents a still later stage where the 
division is complete and two curved xylem strands are produced. In the next section, 
other protoxylem groups make their appearance on the inside of each of the curved 
xylem strands, but they are near the lower ends of these strands and not in the centre. 
Two sections above this last one these extra protoxylems are distinct, and the are of 
xylem is converted into a double are (Pl. ILL. fig. 38, pon. tr., aph. tr.). In Pl. IV. fig. 39 
—ahbout three sections above fig. 38—the small traces are cut off from the larger median 
bundles. The small bundles pass out to supply aphlebiz, and during their passage 
outwards they each bifureate. The two branches of each bifurcation pass out at the 
same level (PI. IV. fig. 40, aph. tr.,', aph. tr.,?), and not, as in Diplolabis romert, at 
different levels. 
The larger median bundles—destined to supply pinnee—become more curved (Pl. IV. 
fig. 41), and finally pass out at the same level into the pinne, of which there are two for 
each pinna-trace-bar. In some cases, however, the incurved ends unite, and the are then 
becomes a closed ring of xylem (Pl. IV. fig. 43). Such a closed trace is very interesting 
and probably indicates an ancestral character. Very few of such annular traces have 
heen discovered, and unfortunately they could not be followed far enough to see if 
they ultimately resumed the open form, though they probably do resume the horse- 
shoe shape higher up. In one secondary pinna-trace a similar character was noted, 
and here the emission of the tertiary pinna-trace is exactly comparable with the 
emission of the secondary pinna-trace from the primary in Clepsydropsis antiqua, as 
shown by Dr P. Berrranp.* 
Apart from these abnormal cases the pinna-trace is open and consists of reticulately 
thickened tracheides with scalariform protoxylem elements in the sinuses formed by 
the incurved ends. The pinne of one order are cut off from the ends of the trace 
of lower order, and pass out alternately at each side. 
The foliage and fructifications of this species are quite unknown, though several 
* P. Bertrand, Btudes sur la fronde des Zygoptéridées, 1909. 
