238 Evans: AIR CHAMBERS OF GRIMALDIA FRAGRANS 
bers are seen to be in three or four layers in the thickened median 
portion of the thallus. As the margins are approached the thallus 
becomes thinner, and the number of layers decreases until only 
the uppermost layer is left. Except in this uppermost layer the 
chambers are usually polygonal in outline and tend to be isodia- 
metric. In the uppermost layer they tend to be elongated verti- 
cally, as shown in the spaces a and 6. That the spaces communi- 
cate with one another is also indicated in the figure. The space 
c, for example, is connected with a space nearer the epidermis, and 
the space d probably represents a passageway to a chamber in 
another section. The figure seems, at first sight, to confirm the 
statements made by Miiller, that both filaments and cell plates 
are present. Immediately beneath the pore there are apparently 
three filamentous outgrowths, e, f, and g, and a plate-like outgrowth 
is clearly shown at h. Of course, as Schiffner intimates, apparent 
filaments may be nothing more than sections of cell plates. In 
the section drawn careful focusing does indeed show that e and f 
are in close contact with another apparent filament in another 
plane, and the same thing is true of other apparent filaments in 
the section. Some of the cell plates, moreover, appear to have a’ 
fluted surface, so that a section cut parallel with the surface of the 
plate might readily give the impression of a series of filaments. 
At the same time there are many apparent filaments which seem 
to be entirely free from one another, and it is impossible to deter- 
mine their true status except by the study of other sections. It 
will be noted that the more deeply situated chambers are free or 
nearly so from outgrowths of any kind. 
The figure is of further interest in showing that some of the 
apparent filaments and plate-like outgrowths end freely in the 
chamber without reaching the epidermis, this being especially 
true in the vicinity of the pores; others, as shown by the one be- 
tween the spaces a and b, extend to the very epidermis and seem 
to be connected withit. It is doubtful, however, if the connection 
is ever anything more than a close contact, such as the free fila- 
ments in Marchantia and Conocephalum often exhibit. No 
instance has been observed where an outgrowth extends downward 
from the epidermis and ends freely in a chamber, and there is no 
adequate evidence that the epidermal cells themselves ever give 
