672 REPORT—1899. 
increase in weight is due almost entirely to the formation of carbohydrates, we 
can calculate with a close approximation to accuracy the corresponding amount of 
carbon dioxide. This will of course depend, within certain narrow limits, on the 
nature of the carbohydrate formed. The formation of a gram of starch requires 
1-628 grams of carbon dioxide, whilst an equal amount of a C,H,,0, or a C,,H,,0,, 
sugar requires 1-466 and 1:543 grams respectively. From the knowledge we possess 
of the nature of the carbohydrates of the leaf, we are quite sure that the mean 
of these values, that is 1°545 grams, must be very near the truth. This amount 
corresponds to 784 c.c. of carbon dioxide at normal temperature and pressure, 
which must represent the volume abstracted by the square metre of leaf 
surface in one hour from air containing only three parts of carbon dioxide in 
10,000, supposing the method of leaf weighing to give correct results. We shall 
see later on that this intake can be verified by direct estimations; it is equivalent 
to the total amount of carbon dioxide in a column of air of a cross section equal 
to that of the leaf, and of a height of 26 decimetres. 
The extraordinary power which an assimilating leaf possesses of abstracting 
carbon dioxide from the air is hest shown by comparing it with an equal area of a 
freely exposed solution of caustic alkali. We have made a very large number of 
experiments on the rate at which atmospheric carbon dioxide can be taken up by a 
solution of caustic soda under varying conditions, and have been surprised to find how 
constant the absorption is. In a moderately still air a square metre of surface 
of such a freely exposed solution will absorb about 1,200 c.c. of carbon dioxide per 
hour, and this can only be increased to about 1,500 c.c. even if the dish is exposed 
to the full influence of a strong wind out in the open. When the surface of the 
liquid is constantly renewed during the experiment by means of a mechanical 
stirrer, the rate of absorption is not sensibly affected, providing the agitation does- 
not appreciably increase the surface area, and considerable variations in the 
strength of the alkaline solution are also without any effect. On the other hand, 
slight variations in the tension of the carbon dioxide of the air have a marked 
influence on the rate of absorption, and in order to study this point we have 
constructed an apparatus which allows us to pass over an absorptive surface of 
liquid a current of air in a stratum of known thickness, and with a known average: 
velocity. 
By trodducing definite amounts of carbon dioxide into this stream of air we 
have been able to determine the influence of its tension on the rate of absorption. 
At present we have only employed air containing amounts varying from 0°8 to 
13 parts per 10,000, that is to say, from about one-quarter to a little more than 
four times the amount contained in normal air. Within these limits, and probably 
beyond them, the rate of absorption by the alkaline surface is strictly proportional 
to the tension of the carbon dioxide in the air current. I shall have occasion to: 
show later on that the same rule holds good with regard to an assimilating leaf, 
and that in this case also, within certain limits, the intake of the gas is propor- 
tional to its tension. 
The fact which I wish more particularly to bring out in these comparisons is 
that a leaf surface which is assimilating at the rate of one gram of carbohydrate 
per square metre per hour is absorbing atmospheric carbon dioxide more than half 
as fast as the same surface would do of wetted with a constantly renewed film of a 
strong solution of caustic alkali. 
From what I have just said about the influence of tension on the absorption of 
carbon dioxide by an assimilating leaf, it is clear that any attempts to determine 
by direct means the natural intake of that gas during assimilation must be made 
with ordinary air, and that such experiments can only be carried out on a com- 
paratively large scale. We had in the first instance to devise an apparatus which 
would rapidly and completely absorb the whole of the carbon dioxide from a 
stream of air passing through it at the rate of from 100 to 200 litres per hour, 
and at the same time admit of an extremely accurate determination of the 
absorbed carbon dioxide. 
The absorbing apparatus which we finally adopted is a modification of one 
used by Reiset in his estimations of the carbon dioxide of the atmosphere. It 
