194 
Gibbs . — Bio-histological notes on some new 
differentiation on each surface of the leaf, those on the dorsal surface being 
much smaller than those on the ventral or upper side. The contents of all 
the cells are colourless, and only the outer walls are thickened, the radial 
and inner tangential walls being of an extremely delicate texture. 
On the dorsal surface they are smaller over the hypodernial ribs 
(PI. XII, Figs. 12 and 13), where the development of the hard fibrous cells 
seems to arrest their radial expansion, which is shown in a transverse sec- 
tion through a developing leaf of C. stellulata Good (Fig. 15). In the 
youngest stage obtainable, the sclerenchymatous cells not having yet 
differentiated out, those of the epidermis are still more or less equal in 
diameter. The peculiar cone-like base limited to the most radial of these 
cells noted by Duval-Jouve 1 in F. pubescens Poir. was not seen in the 
species examined. The ventral epidermis is marked by enormous thin- 
walled cells with colourless contents (PI. XII, Figs. 12 and 13). The outer 
walls alone are cuticularized together with the unicellular hairs which are 
scattered over its surface, the radial and inner tangential walls are very 
delicate in texture. These cells in the vicinity of the vascular strands are 
radially elongated, and there is always an intimate connexion with the 
bundle-sheath of delicate thin-walled cells (Figs. 12 and 13). This con- 
nexion is continuous with the dorsal epidermis by means of small colourless 
cells passing up both sides of the sclerenchymatous ribs which accompany 
the bundles on that side. 
The large cells which occur in groups in the epidermis of grasses, and 
throughout the Cyperaceae, characterize the entire ventral epidermis of the 
leaf in varying degrees of differentiation, and have been described as ‘ bulbi- 
formes’ by Duval-Jouve, who observed that stomata were always absent on 
a f bulbiforme ’ epidermis. Tschirch 2 , who looked upon them as character- 
istic of steppe and meadow grasses, gave them the name of c Gelenkzellen ’, 
as he considered they controlled the folding and unfolding of the leaves of 
grasses by their turgescence or loss of water. Volkens 3 states that he has never 
observed the folding up of the lamina of a grass leaf, but always a windbag 
(Blasebalg) action, as the lamina increases or decreases in width according to 
the turgescence (Wasserzufuhr) of these cells, which proves they function as 
a water reserve. He therefore rechristened them ‘ Wasserzellen ’. 
Rickli 4 , noticing the delicate radial and inner tangential walls of all the 
epidermal cells, looks upon the whole organization as a water-jacket, and 
there can be no doubt that this interpretation is correct. This is shown by 
the connexion, between the two surfaces, with the bundle-sheaths, as already 
pointed out, by the delicate inner walls, and by the limitation of the stomata 
1 Duval-Jouve, Sur une forme de cellules epidermiques qui paraissent propres aux Cyperacees 
(Bull. Soc. Bot. France, xx, 1873, pp. 91-95). 
2 Tschirch, Beitrage zu der Anatomie u. dem Einrollungsmechanismus einiger Grasblatter 
(Pringsh. Jahrb., xiii, 1882, pp. 544-68, Pis. XVI-XVII 1 ). 
3 Volkens, Flora der agyptisch-arabischen Wiiste, 1887. 
4 Rickli, 1 . c. 
