Notes . 
165 
parative readiness with which C 0 2 osmoses through thin films of 
caoutchouc, with observations by Fr^my and others on the similarity 
between cuticle and caoutchouc in chemical composition, first led 
Barthelemy (1868) to put forward the view that the cuticle was 
specially adapted for transmitting C 0 2 from the external air to the 
assimilating cells beneath it. This view he supported by experi- 
ments on the artificial osmosis of gases through leaves. About the 
same time Boussingault performed experiments that seemed to 
definitely show that in assimilation* the C 0 2 taken up by the leaf 
entered it through the upper surface, devoid of stomata, to which 
the assimilating cells are adjacent, rather than through the more 
distant stomatal openings. These experiments have hitherto been 
generally accepted, but I shall show later that the conclusions drawn 
from them are entirely fallacious. In support of the view that 
stomata form the paths of gaseous exchange, besides isolated induc- 
tions by various workers, we have the conclusion arrived at in 1888 
by Mangin, from diffusion experiments on isolated cuticle, that this 
diffusion is insufficient to account for the whole gaseous exchange of 
the leaf. 
By the aid of the apparatus described in a previous paper the 
author has been able to successfully attack the problem directly by 
estimating the amounts of C 0 2 given out or taken in by the two 
surfaces of the same leaf, under the same conditions. For this pur- 
pose shallow capsules, 10 sq. cm. in area, consisting of a glass plate 
with a metal rim, through which tubes for the circulation of the air 
current pass, are employed. Two of them are affixed to a leaf on 
opposite sides of the same area in air-tight union by means of soft 
wax. Then, in the way described previously, two continuous currents 
of air can be kept up over the two surfaces, and the C 0 2 produced 
or taken in during a given time by each of them be determined. 
Numerous experiments on the respiration of a variety of leaves, 
thick and thin, with the stomata all on the one side or with stomata 
variously distributed on the two sides, agree in showing that the 
stomata are the site of the exhalation of this gas. When no stomata 
are present on the upper surface of a leaf then practically no C 0 2 is 
exhaled from that surface, at least not more than falls within the 
working error of the apparatus under these conditions, while more 
than thirty times as much may be given off from the lower stomati- 
ferous surface. When stomata occur on both surfaces the relative 
