410 DR. OTTO STAPF ON THE 
There would therefore be a considerable gap in my account of the development of the 
fruit, but for the light which certain remarkable young fruits from the Garo Hills threw 
on the earlier phases. These fruits ranged from the earliest stages to that of almost 
complete maturity. They (Pl. 45. figs. 16-18) were mostly normal to all outward 
appearance, but never contained a trace of an embryo, representing evidently a case of 
imperfectly efficient fertilization, similar to that of seedless raisins, citrons, hops, or 
Crucifers +. Apart from the absence of an embryo and certain changes in the apex 
of the ovule immediately dependent on the presence of the embryo, the development of 
the fruits was no doubt essentially that of normal fertile fruits. The changes in the apex 
of the ovule just alluded to consist in its being crushed and obliterated by the growing 
embryo, which, as it enlarges, fills the bottom of the ovarial cavity and pushes its way 
towards the base of the fruit. No embryo having been formed in the barren Garo Hills 
fruits, the apex of the ovule took part in the general enlargement of the ovule until in one 
case it attained the considerable length of 4 mm. (Pl. 45. figs. 18, 19). Making, then, 
use of my observations on these fruits for the description of the earlier stages of the 
development of the fruit, I proceed now to give as far as possible a coherent account of it. 
The ovule grows, after fertilization, simultaneously with the ovary. The ovarial 
cavity also expands, but chiefly longitudinally, so as to form a flat linear canal (fig. 16, oc.*) 
in front of the embryo-sae. Yet, as the latter grows faster than the cavity expands, this 
canal is not only left behind in its longitudinal development, but is also gradually closed 
up, the obliteration beginning from the upper end. It remains, however, for a long 
time distinctly discernible as a narrow marked strip, situated along or somewhat aside of 
the middle line of the front half of the embryo-sac, and ascending to about 3 to 2 of its 
length. As the ovarial cavity closes up, the meeting surface layers of the fr ont e back 
walls become more or less liquefied or jellified. How early the peripheral cells of the 
ovule-wall prepare for this change, I have already stated on p. 409. In consequence of 
this process the ovarial canal may be filled up with a structureless substance (Pl. 45. 
figs. 20 oc.*, 21); but as a rule the walls coalesce without leaving a trace of their original 
separation, or only betraying it, in cross-sections, by somewhat thin bands with the 
appearance of solidified jelly. Whilst this is going on, endosperm (Pl. 45. figs. 20, 22, end.) 
forms as a delicate film, lining the embryo-sac with a layer of small almost isodiametric 
cells, which are followed inwards by larger radially lengthened cells (Pl. 46, figs. 30, 31). 
The thickness of the lining is, however, more or less unequal in different parts, so that 
its inner surface is very uneven (Pl. 45. figs. 20, 22, end.). Occasionally flake-like 
extensions of endospermatie tissue project far into, or even traverse, as thin strands or 
lamellee, the cavity, which is filled with an almost limpid, slightly opalescent liquid. 
Apart from such exceptional expansions, the endospermatic lining grows rarely thicker 
than 1 mm., and often remains considerably thinner. The cell-walls of the endosperm 
are eg d thin, Vesdegerege those of the inner portions, which in some cases were found 
E + See Noll, & « Fruchtbildung ohne vorausgegangene Bestüubung (Parthenoearpie) bei der Gurke,” in Sitz.-Ber. 
= hein. Gesellsch. für Natur- u. Heilkunde zu Bonn, 1902, p. 1. True parthenocarpy is, in my opinion, out of 
- question, v that the Melocanna forests of the Garo Hills must have been full of pollen. 
