146 DR. F. E. FRITSCH ON THE 
isodiametric shape, so that they constitute a kind of inner epidermis. The wall of the 
involucre has very much the same structure as the outer portion of the ovary-walls, 
being composed of thin-walled parenchymatous tissue traversed by resin-canals and 
vascular strands. The outermost layers are brown in colour and contain numerous 
clustered crystals of medium size; hairs occur sparingly on the outer surface of the 
involucre. 
The structure of the mature fruit is far more complicated, but is very nearly 
uniform in all the cases investigated ; so it seems advisable to first describe the anatomy 
for one species and then to mention the points of difference presented by the others. 
A transverse section through one of the walls of the original ovaries in thé ripe fruit of 
J. amplifolia (Rose, 3735) or J. adstringens (Pl. 21. figs. 1 & 2) shows the following 
structure :—The cavity of the ovary, which contains a single seed, is lined by a layer of 
cells, which are cubical to rounded or of irregular shape in transverse seetion, and have 
very strongly thickened and often distinctly stratified walls (Pl. 21. figs. 1 & 2, i.ep.). 
In most cases the thickening goes so far that only a small dot-like lumen remains, from 
which very fine canals frequently radiate outwards. These cells have a thin cuticle and 
form a perfectly continuous and smooth layer lining the cavity of the ovary internally. 
In surface-view (Pl. 21. fig. 3, i.ep.) the cells are seen to have a more or less undulating 
outline, and the way in which the lumen commonly radiates out into the thickening 
layers of the walls is readily seen. Both in general shape and position these cells agree 
with those forming the inner limit of the wall of the ovarial cavity in the young 
fruit (cf. above), and they undoubtedly arise from the latter by the apposition of 
thickening-layers to the original cell-wall. Immediately external to this inner lining 
layer we meet with the same characteristic crystal-containing cells as in the young 
fruit (Pl. 21. figs. 1 & 2, c./. 1); these cells generally still have a more or less plano- 
convex shape, but some have become rounded or ellipsoidal in transverse section. Their 
walls are considerably thickened and provided with a small number of simple pits; 
a well-developed lumen remains, however, and this is praetically filled out by a large 
solitary erystal of oxalate of lime. In a surface-view of the ovary-wall these cells 
appear very large and rounded (Pl. 21. fig. 3, c./. 1), the crystals contained. in their 
lumen often having rather an irregular shape. These cells, with their large crystals, 
form a very pronounced although interrupted line immediately beneath the inner 
epidermis of the ovary-wall; they absolutely correspond to the layer occupying the same 
position in the young fruit. The remaining structure of the ripe fruit is not so easily 
correlated with that of the young one. ‘The line of subepidermal erystal-containing cells 
is not continuous, but is interrupted by occasional small cells (Pl. 21. fig. 1) without 
erystalline contents and having only a very small dot-like lumen. These cells really 
belong to the mass of tissue lying external to the crystalline layer; this tissue (Pl. 21. 
figs. 1 & 2, m.) consists of several layers of thick-walled cells, and has developed in the 
position formerly occupied by the thin parenchyma, traversed by the vascular bundles; 
of the latter no traces are to be found in the mature fruit. The cells of the tissue 
in question are fairly large and have strongly thickened more or less distinctly stratified 
walls, the thickening being so extensive that the cavity of the cell is in most cases 
| 
1 
| 
| 
| 
| 
