an archaic type of Seed from the Palaeozoic Rocks. 83 
It will be seen from the foregoing general sketch of the organization 
of Physostoma that, in its fundamental features, it is modelled on such a seed 
as Lagenostoma Lomaxii. Like Lageno stoma, Physostoma has nucellus and 
integument coalescent throughout the greater part of their extent ; at the 
apex only, where the nucellus ends in a pollen-chamber, are these organs 
free. The relations of the vascular supply to the integument are identical 
in the two cases, and the pollen-chambers have much in common. 
The outstanding features of difference in the gross structure of the two 
seeds are to be found in the free portion of the integument. In Lagenostoma 
this consists of nine arcs joined together to form a chambered dome — the 
‘ canopy ’ of Williamson — through the midst of which the tubular orifice of 
the pollen-chamber projects to the surface. In Physostoma , on the other 
hand, the canopy is represented by the whorl of free arms ; the spatial 
relations are the same, but fusion is lacking. It is evidently quite inappro- 
priate to apply the term micropyle here, a term designed for an integumental 
tube leading down to a nucellus and giving access either to pollen-grains 
or to a pollen-tube. This remark applies with almost equal force to the 
case of Lagenostoma , where the orifice of the pollen-chamber is produced 
beyond the surrounding canopy, and must have received the pollen-grains 
without any assistance from that structure. In dealing with the seeds of 
such remote ages, the possibility cannot be ignored that features may be 
present belonging to a period anterior to the evolution of what is termed 
a micropyle (cf. Text-figs. 1, 2, and 10). 
The main point to be recognized here is that the tentacles of 
Physostoma represent the units of the L agenostoma-cano^y in an unfused 
and, it may well be, in a more primitive condition. 
Another point may be mentioned here. Lagenostoma Lomaxii is 
occasionally found still enclosed in a free, lobed sheath or cupule — like 
a hazel-nut in its husk. No structure of this kind has been detected in 
any of the specimens of Physostoma elegans , though the existence of 
a cupule is not of course excluded. Lagenostoma , with its fairly broad 
seed-base, had a very effective abscission layer and was shed quite early ; 
Physostoma , with its tapering base, may well have been even more 
readily detached. 
3. Numerical variation. 
During the course of the present investigation, sections of about fifty 
specimens of Physostoma have come under observation in which it was 
possible to count the number of ribs or tentacles. 
The result is as follows: Seeds with 9 ribs, 8; 10 ribs, 25; 11 ribs, 
14; 12 ribs, 6. Total seeds counted. 53. Two seeds apparently showing 
8 ribs were rejected as doubtful. 
G 2 
