73 
Daisy G. Scott. 
on the stamens, springing from the tissue facing the axis of the 
flower. They covered the whole available surface on this side of 
the stamen, and were scattered about without any apparent order. 
The transverse sections generally cut the ovule longitudinally, thus 
allowing the embryo-sac and integuments to be clearly seen. Often 
however the ovules did not grow out at right angles to the axis of 
the stamens, but were directed obliquely upwards, (Fig. 11, I). 
Another curious point in those sections was that many ovules cut 
in various planes appeared to lie freely in the space between the 
ovary and the stamens. This effect is explained by the fact that 
just below this level the stamens suddenly bulge out to form a ridge. 
This ridge, like the whole interior face of the stamen at this level, 
was covered with ovules, many of which shewed no connexion in 
transverse section with the stamen on which they grew. 
In the flower under consideration four of the stamens shewed 
the peculiarities described, the fifth producing no ovule on any part. 
It exhibited one peculiarity however, namely, that it was partially 
petaloid (II), being flattened out very much laterally. The 
anther lobes were no longer distinct from the connective, and the 
whole structure appeared in transverse section the same thickness 
throughout. At both ends pollen-sacs were produced, hut they 
were much smaller than the pollen-sacs of other stamens of the 
same flower at the same level (cf. I and II). At one end of the 
stamen there was one pollen-sac which would probably rupture 
on the inner side, since the parietal tissue was thinnest there. 
There was no indication of a second pollen-sac. The other end 
gave by its curved outline at any rate some suggestion that two 
pollen-sacs might have formed one loculus. The sac in this case 
would rupture on the inner side of the stamen. 
Transverse sections were made of a second and somewhat 
older flower, and it was found that in this case there were five 
stamens arranged alternately with the petals, but these stamens 
were much less regular in shape than those in the first flower 
examined. The ovary in this, as in the other flowers, was normal. 
Sections taken at a fairly high level shewed stamens each with an 
irregular outline and producing pollen-sacs, but not always in the 
situations one would expect. Sections at a lower level shewed the 
appearance represented diagramatically in VI. The stamen on 
the right hand was very similar to those found in the first flower 
and produced ovules freely from its inner surface. The typical 
outline, however, was less well maintained, and only one pollen-sac 
