AND AFFINITIES OF STEPHANOSPERMUM. 365 
layer, which may be termed the epithelioid layer of the sclerotesta, becomes continuous 
at the chalaza with the epidermis of the nucellus, and agrees with it in its histological 
characters (Pl. 43. fig. 21, e£. & ne.). All the layers of the sclerotesta appear thick- 
walled and sclerotic, and all but the lining layer have brown cell-walls. "The walls and 
contents of the former are often excellently preserved. The protoplasm is contracted 
away from the wall except at certain points where finger-like processes remained, as in 
fig. 22, which shows the ends of two sclerotic cells from the third layer from the surface. 
Fig. 23 represents another cell from immediately below the palisade-layer. In almost all 
the cells the nucleus appears to be represented by a dense body usually surrounded 
by a light contour. In the palisade-layer the preserved contents take the form of a long 
and often slightly sinuous filament with a local enlargement where the nucleus is 
contained (fig. 24, n., also fig. 21, /.p.). At the peripheral ends of these cells, with rare 
exceptions, and sometimes at their internal ends, the filament expands into a low, hollow, 
conical vesicle, the base of which is closely pressed against the limiting membrane (as at 
v. fig. 24). More frequently we find at the internal ends paired finger-like processes 
(p., fig. 24) which vividly recall pits. Occasionally, in the same position, there is a vesicle 
with a pair of short processes directed towards the cell corners. No doubt all these 
appearances were occasioned by the adhesion of the protoplasm to the wall, and it is 
difficult to resist the inference that a pitting of the end walls has been largely concerned 
in the production of the finger-like processes. But a detailed discussion of the matter 
would not be appropriate here.  Palisade-cells are shown in transverse section in fig. 25. 
The apparent great thickness of the walls of the palisade and supporting cells of the 
sclerotesta is very probably exaggerated through preservation. The direction followed 
by the former is horizontal at the middle of the seed (where it is broadest) ; above this 
zone, and below, these cells gradually acquire an increasingly steep angle, till at the chalaza 
their course is parallel to the axis of the seed; whilst at the “crown” or perimicropylar 
ring they run up directly into this structure (cf. Pl. 41. figs. 1 4 & B). The apex of the 
seed merits detailed description. The sharp ridge or crown which surrounds the 
micropylar beak is composed of an extension of the palisade-cells. The trough or 
moat which it contains is generally empty (figs. 14 & 5). Occasionally, however, the 
lining-surface of the trough bears a loose tangled feltwork of long filaments forming a 
brush inserted over the concave surface and reaching up to the micropyle (Pl. 41. figs. 2, 5, 
& Pl. 42. fig. 10). The component parts of the feltwork show little obvious indications 
of cellular structure. Two sorts of elements can be distinguished. First, there are 
indistinct remains of tubular cell-membranes much crumpled and distorted, and, mixed 
with these, filaments of extreme fineness, often following a sinuous course and occasionally 
expanding into small lenticular thickenings. Taken in conjunction with the normal type 
of preservation of the palisade-cells just described, the matted feltwork becomes 
intelligible. The membranes are the much-macerated walls, and the sinuous filaments 
the contracted contents of long cells or rows of cells. In some cases the membranous 
portion has been denuded away, and the tissue which once occupied the trough is 
represented by its protoplasmic contents alone. Of the original nature of this tissue it 
is difficult to judge. It may have consisted of tufts of real hairs, or of a spongy 
