AND AFFINITIES OF STEPHANOSPERMUM. 375 
well-defined and symmetrical, its limits being marked out by the remains of its broken- 
down cells which had become absorbed in the excavation of the cavity. It is in this 
peripheral mucilage, especially, that the pollen-grains are found preserved. Their 
position, in a number of cases, is shown in text-fig. 2, p., p°, &c., whilst some are visible 
in Pl. 42, fig. 18. To this matter we shall return on p. 379. Of the cleft or gap below the 
pollen-chamber, so conspicuous a feature in S. akenioides, relatively small trace is found. 
A certain amount of splitting there has been, and the remains of the septum may be seen 
at s. (Pl. 42. fig. 18); the smaller photograph (fig. 17), with a hand-lens, shows it on the 
right-hand side as well (see also text-fig. 2, s.a.g.). The splitting is of the same nature 
as in AN. akenioides, but less in degree. The floor of the supra-archegonial gap is 
practically destitute of tracheides, as are the sloping sides of this space, but the significa- 
tion of this point of difference will be considered below. The main portion of the 
nucellus is, as usual in seeds of this period, reduced to a sheath or pouch enclosing the 
large macrospore. Similarly, an epidermis—well-defined, though at places detached 
from the nucellus—is present; at the chalaza it becomes continuous with the epithelioid 
lining the sclerotesta (text-fig. 2). As stated above, though the entering bundle is not 
included in any of the sections, there is no reason for supposing that it differed from 
that of S. akenioides. In the chalazal region of the nucellus there are ample remains of 
au extensive tracheal plate, the margins of which are continued up the nucellus as 
an unbroken sheath of tracheides, traceable up to the insertion of the sides of the pollen- 
chamber (text-fig. 2, b). The layer of tracheides lies just within the surface of the 
nucellus ; between it and the epidermis are the compressed remains of a soft peripheral 
layer. In other words, the relations here are essentially as in the allied species. 
At one spot near the chalaza a large fragment of the nucellar wall has splintered 
away and been drawn into the cavity of the embryo-sac (Pl. 42. fig. 17, f). "This is 
repeated in all three of the more median sections and is not due to shattering. ‘The 
fragment of the macrospore and nucellar wall was evidently displaced anterior to 
mineralisation and in this position it has become embedded. The tracheides of this 
splinter have become somewhat fretted out and are in some degree isolated. They are 
unusually large and well preserved, and form most beautiful mieroscopie objects. 
Pl. 44. fig. 39 shows a group, whilst two of the tracheides are represented in Pl. 42. fig. 19. 
The rib-like thickenings of the walls show an unusual perfection of preservation. 
Occasionally the scalariform type gives place to a reticulated thickening as in fig. 39. 
The average diameter of these cells is 20 4, so that they are much larger than in 
S. akenioides (where the diameter is about 14,4). This type of tracheide forms a 
continuous sheath of closely packed elements in the nucellar wall, some three deep, 
traceable to the insertion of the pollen-chamber, where they die out (text-fig. 2, b). 
It may be conjectured that on the top of the embryo-sae they have been absorbed, 
a change which, apparently, likewise took place in S. akenioides (cf. p. 370), where in 
older seeds they are sometimes less easy of observation and occasionally wanting 
altogether. That the solitary specimen of S. caryoides corresponds to a stage slightly 
older than the commonly occurring type of Stephanospermum there is good reason to 
suppose (cf. below, p. 387). ` 
