gl The Anatomy of Ginkgo biloba. 
directed towards what would be the upper surface. The relationship 
is however obscured owing to the. excessive development of one 
ovule, which comes to lie almost apically. The pedicel passes very 
rapidly into the collar with its inverted bundles. 
Now in the seed Lagenostoma Lomaxi it is stated that the 
pedicel has the structure of a petiole, and, what from our point of 
view is more important, in the cupule the vascular bundles are often 
inverted (Oliver and Scott Phil. Trans., Vol. 197 B, 1904). It is 
only fair to state however that the general impression in the minds 
of the authors seems to have been that the bundles were, as a rule, 
normally orientated although there was a considerable amount of 
variation. At all events for the present we would tentatively 
suggest that, in the light of anatomical investigation, the collar of 
the ovule in Ginkgo is better regarded as a vestigial cupule than as 
anything in the nature of a carpel. Perhaps when our knowledge 
of the Pteridosperm seeds with cupules is more extended we may 
find some in which the cupular bundles were, as a rule, inversely 
orientated. 
In this connection it is interesting to note that at the base of 
the seeds of certain Cycads (C. revoluta, C. Rumphii) we have a 
state of affairs closely paralleled in the collar of Ginkgo, the single 
vascular bundle supplying the ovule being accompanied on its outer 
convex side by a number of inversely orientated strands (Worsdell, 
Annals of Bot. XII., 1898, PI. XVII., Fig. 4). It seems therefore as if 
both Ginkgo and the Cycads might have been derived from forms with 
cupuliferous seeds, a conclusion which coincides with other evidence. 
By some writers the collar has been considered as an outer 
integument, but if we are to look for an outer integument in the 
lower Gymnosperms we are much more likely to find it in the double 
nature of the real integument of Ginkgo and the Cycads than in the 
“ collar” of the former, which appears to be essentially a “cupule,” 
an organ sui generis. 
Of course it may be urged that the vascular bundles described 
in the present paper simply indicate that Ginkgo originally possessed 
a vascular structure similar to that of the Medulloseae. This view 
however would take no account of the peculiar situation of the 
anomalous tissue and, further, it offers no explanation of the “ collar.” 
It should also be remembered that the seed is a large one, and 
this fact, taken in conjunction with the large development of 
“ transfusion tracheids,” renders it possible that the presence of 
additional vascular tissue may have a physiological, rather than a 
morphological, significance. 
