an archaic type of Seed from the Palaeozoic Rocks. 105 
In view of these considerations it is suggested that the integument of 
Physostoma may be an example of encasement, a structure, that is to say, 
whose origin is contemporary with the inception of the seed habit. Here, 
as in other members of the group, it has become coalescent with the central 
sporangium. Those seeds which possess cupules, like Lagenostoma Lomaxii 
and L. Sinclairii , according to this view, have undergone a second encase- 
ment, though the envelope is much less specialized than in the case of the 
integument proper. In later groups, such as the Angiosperms where, in 
addition to two ‘ normal ’ integuments and a carpel, an aril is often present, 
we have further phases of the same function. 
Assuming the view to be well-founded that the integument of the seeds 
of the Lagenostoma-group was a lobed structure at its inception, it is not 
without interest to see how far this primitive feature has impressed itself 
upon the several seeds. 
In Physostoma , as we have seen, the tentacles and ribs are the 
conspicuous feature at the surface, whilst deeper down the vascular strands 
are in perfect correspondence. In addition to the integument the tissues of 
the nucellus in the body of the seed are slightly involved, for the secretory 
zone below the ribs shows more numerous ranks of secretory sacs than 
occur between the ribs. 
Conostoma oblongnm stands next to Physostoma. The free part of the 
integument is united around the pollen-chamber, it is true, but the unit- 
portions show a distinct tendency to separate, and the lines of fusion are 
often marked by shallow grooves. The body of the seed is smooth and 
circular in section and without trace of ribbing save at the chalaza, where 
sharp-angled ridges appear overlying the vascular strands. 
In Lagenostoma Lomaxii and L. ovoides the free part of the integument 
is perfectly united, and the body of the seed smooth throughout. The 
multiple origin of the integument, however, is well seen in the chambered, 
free integument — Williamson’s ‘canopy’ — each chamber representing the 
apex of an original integumental lobe. The vascular strands correspond 
in number with the chambers of the canopy, though in very occasional 
specimens their number may be reduced by fusion lower down the seed- 
wall. The little ridges which surround the micropyle, and are conspicuous 
in some sections — especially in L. Lomaxii — should not be mistaken for the 
unit-lobes of the integument. They overlie the well-marked septa between 
the chambers of the canopy, and it is quite possible they owe their 
prominence to post-mortem contraction or collapse of the filling-tissues of 
the chambers. 
The course pursued by the vascular strands at the chalaza may be 
recalled here. Whilst in the Lagenostomas a common supply bundle runs 
up almost to the embryo-sac before the peripheral strands separate out, 
