462 MR. J. R. GREEN ON THE ORGANS OF 
in different plants. Martinet describes* similar structures in the 
ovary of Dictamnus Fraxinella which are borne on processes of 
the carpellary leaves. These, he says, consist of an epidermis 
enveloping a mass of cellular tissue, in the midst of which is a 
spiral vessel running to the gland but not entering it. The 
glands of Drosera rotundifolia, described by Darwin t, show a 
similar relationship, as do the water-glands of Saxifraga incrus- 
fata $. Their development is most easily traced in the very young 
sepals of Hypericum hirsutum. The youngest tooth of such a 
sepal is drawn in fig. 15, Pl. XX XIX. : it consists of an epidermis 
clothing a mass of cellular tissue. The cells of all this mass are 
very delicate and transparent. One cell, just underneath the epi- 
dermis at the apex of the tooth, has a darker colour and its con- 
tents are slightly granular. From this cell the gland arises by 
divisions in irregular order (fig. 16). As the tooth grows the 
cells behind the gland multiply faster than those of the gland, 
and the latter is forced farther and farther outwards. While it 
is still developing, the tissue in the centre of the tooth behind 
the gland becomes differentiated and a fibrovascular bundle 
makes its appearance, forming a branch of the main fibrovascular 
system of the leaf. The cells of the external part of the spherical 
mass of tissue are somewhat flatter and narrower than the in- 
terior ones. The whole outer part of the gland is covered 
directly by the epidermis, which forms a kind of bounding layer 
to it (figs. 15, 16, 17). The cells do not undergo absorption as 
do those of the pellueid reservoirs, but persist through the 
gland's whole life. All through the time of their existence they 
are dark in colour, becoming darker, however, as they mature. 
There is, as a consequence of their persistence, never any cavity 
containing secretion, the latter continuing in the cells. 
In H. perforatum ihe glands are not situated with the same 
regularity on the leaf. Generally a row of them is found round 
the edge, with several others irregularly scattered about it, 
especially near its apex. The same relationship with the fibro- 
vascular bundles is maintained. The fibrovascular system is 
arranged in very fine reticulations, and branches of the ultimate 
network may be traced to every gland. In some cases they do 
not appear to end at the gland, but rather to pass close beneath 
or above it. In other cases the ending can be distinctly seen 
* Loc. cit. t ‘Insectivorous Plants’ (London, 1875), p. 6. 
1 Gardiner, Quart. Journ. Micr. Sci. 1881, p. 407. 
