STRUCTURE OF BAMBOO LEAVES, 19 
Pée-Laby, in Ann. des Sc. Nat. 8"* sér, (1898), viii. p. 314, calls the cells of the upper 
epidermis, described as cellules bulliformes by Duval-Jouve, cellules motrices; Marshall 
Ward (1901) and Lewton-Brain (1903) call them  motor-cells. Hackel (1882) and 
Theodore Holm (1891-1896) retain the name given by Duval-Jouve, and following these 
authors I shall speak of them as bulliform cells. 
In Bamboos, although their development certainly enables the young rolled-up leaf to 
spread out flat, these cells when mature, in a large number of species, become filled with 
solid silica entirely or partially. The effect is to make the blade of the mature Bamboo 
leaf rigid and not motile. This remarkable fact induced me when I first discovered 
it, several years ago, to call them silica-cells, which term I adopted on p. 661 of my 
‘Indian Trees. Since that time, however, I have found so many species with few only of 
these epidermis-cells filled with silica, that I now prefer the designation bulliform cells. 
These siliea-bodies are colourless; when calcinated they remain unchanged, and they 
are not affected by nitrie acid. "They are not crystalline; under the polariseope they 
have the appearance of opal. They can readily be distinguished in a piece of leaf that 
has been boiled in water and has afterwards been placed in a concentrated solution of 
phenol (10 parts of water to 100 parts of crystallized phenol) In some species ( Cephalo- 
stachyum, Pl. 13. fig. 21, Ochlandra) nearly all the bulliform cells are filled with silica, in 
others (Arundinaria falcata, Pl. 11. fig. 1; Melocanna bambusoides, Pl. 13. fig. 27) about 
one-half of the cells are empty, and in some species (Dendrocalamus giganteus, Munro, 
and D. Brandisii, Kurz) the silica-bodies are scanty and at times entirely absent. 
Young leaves while rolled up have no silica-bodies in the bulliform cells, and it will 
be a matter of considerable interest to study the history of their development as the leaf 
matures. "The walls of the bulliform cells as a rule are thin, and with chlor-zinc-iodine 
show the reaction of cellulose. In some cases the outer wall is thick; the substance of 
the thickening requires further enquiry. Longitudinal sections of leaves, passing through 
the band of bulliform cells, are represented in PI. 11. fig. 10, Bambusa vulgaris, Schrader, 
and Pl. 13. fig. 25, Dinochlora andamanica, Kurz. 
BANDS or BULLIFORM CELLS IN THE LEAVES OF OTHER GRASSES. 
A large number, perhaps the majority, of Grasses have, like Bamboos, bands of bulli- 
form cells alternating with the longitudinal nerves on the upper epidermis only. ‘There 
are, however, other modes in which these remarkable cells are distributed, and it may be 
useful briefly to review them, following in the main points the classification established 
by Duval-Jouve (* Histotaxie, p. 318). 
A. Bands of bulliform cells in the upper epidermis only. 
a. Alternating with longitudinal nerves. 
I. Panicez. Panicum (partly), Pennisetum (partly), Spartina. 
AIL. Oryzee. Oryza. ; 
V. Zoysieæ. Tragus. 
VI. Andropogonez. Saccharum, Elionurus, Andropogon. 
VII. Phalarideæ. Phalaris (incl. Digraphis) (partly), Hierochloa, Crypsis, Alopecurus. 
