1904] BILLINGS—TILLANDSIA USNEOIDES 117 
outgrowths of the parenchyma. The outgrowths do not stop, but 
merely hinder transpiration. Xanthorrhoea hastilis exhibits a similar 
contrivance. Camellia japonica and Prunus Laurocerasus have the 
faculty of filling up the air space as a result of excessive drouth or by 
death of the guard cells. In such cases tylose-like processes occur 
which block up all gas interchange. Pilea elegans differs from those 
mentioned above in that certain subjacent parenchyma cells develop 
thickenings on their exterior walls. One of these finally pushes up 
against the pore of the stoma and effectually closes it. There is no 
movement of the parenchyma cell away from the stoma, hence the 
aperture is permanently closed. From an examination of jigs. 72 
and 73 it will be apparent that Tillandsia presents a condition of 
affairs not widely different from that of Pilea. The principal differ- 
ence lies in the fact that in Tillandsia the parenchyma cells undergo 
no thickening. Both longitudinal and cross sections through the 
leaf show outgrowths from the parenchyma cells lining the sides of 
the air space. The outgrowths turn upward and either stop up the 
opening of the stoma or else press directly against the guard cells. 
It will be seen that the enormously thickened walls of the guard cells 
preclude a possibility of change in their form. To show this experi- 
mentally some plants were placed in water and exposed to direct 
sunlight for a few hours. The leaves were then sectioned and the 
guard cells watched with a micrometer while glycerin was run under 
the cover glass. There was no measurable change. According to 
Mez (6) the guard cells have lost the power of functioning, this power 
having been transferred to certain cells of the subjacent tissue which 
operate the passive guard cells, thus opening and closing the stoma. 
There are two cells which come in contact with the guard cell and 
may therefore be the means of moving it. One is the cell to which 
it is attached and which extends from the hinge to the inner face of 
the guard cell. This cell is usually continuous, but may be divided 
by a cross wall into two cells. Should this cell, which is epidermal, 
become turgescent, it would tend to raise the guard cell, swinging its 
free side outwards. Such a movement, however, would close rather 
than open the pore of the stoma. The hinge is quite thick and may 
be much thicker than any shown in the figures. If the epidermal cell 
is divided the division wall would effectually hinder any movement of 
