g6 Oliver . — On Phy so stoma elegans , Williamson , 
of the ‘ canopy 5 (cf. p. 109). 1 Owing to the contraction of the zone of 
cells along which they run, the bundles in Lagenostoma are very rarely found 
in situ ; commonly they occur attached to a loose sheath which lies about 
midway between the surface and the axis of the seed. In Phy so stoma no 
such displacement is found ; the bundles lie in a zone at the inner limit 
of the tissue of the integument where it abuts on the zone of secretory sacs 
(nucellus). 
Another characteristic feature of our seed is afforded by the course 
followed by the individual strands as they pass from the chalaza, Instead 
of arising from the chalazal bundle, relatively high up and near the base of 
the embryo-sac, these strands are already recognizable at a point not far 
removed from the actual seed-base. No critical section has come under 
observation to settle the point whether the vascular supply entered the seed 
as a single strand, but there are several which show that a ring of contiguous 
strands surrounding a £ pith ’ was present very close to the abscission-zone 
(PL VI, Figs. 19, so, and 21). These strands, usually ten in number, 
gradually diverge and make their way to the apex of the seed at the 
inner limit of the integument. The funnel-shaped ‘ pith 9 below the 
embryo-sac was filled with secretory sacs (see p. 97). At the summit of 
the seed the strands pass out into the tentacles — one to each. 
A great feature in this seed is the system of lacunae that constantly 
accompanies the strands. At the base, the divergent ring of bundles is sur- 
rounded by a continuous annular chink or lacuna (PI. V, Fig. 1). From \ to f 
of a millim. higher up — where the bundles have reached their full peripheral 
extension — a single lacuna for each strand replaces this common ring 
(Fig. 2). These spaces, which are semi-circular or crescent-shaped, lie 
immediately outside the bundles, and pass with them into the tentacles. 
They probably originate from post-mortem contraction or decay. 
Histologically, the bundles consist of xylem only : no phloem elements 
have been detected, but it is possible the phloem may have broken down to 
form the lacunae just described. The xylem strands are extremely delicate 
and rarely show more than six tracheal elements in any transverse section 
(PI. V, Fig. 6 ; PI. VII, Fig. 28). The largest tracheal elements do not exceed 
10-12 \x in diameter, except at the chalaza, where they are somewhat tub- 
shaped, whilst 50 jut x 30 j jjl are usual dimensions for the strands. The 
smallest elements generally occur at or near the outside, so that it is pro- 
bably correct to describe the xylem-strands as exarch or mesarch. The 
tracheal elements are for the most part of the delicate, scalariform type 
(PI. VII, Fig. 27), though traces of pitted elements have been seen in the 
chalazal region, An occasional fine spiral element has been found, but the 
relation of these to the rest of the wood has not been ascertained. 
The bundles, after they enter the tentacles, become much attenuated, 
1 Oliver and Scott, On Lagenostoma Lomaxii, p. 20S. 
