TT. 
TRANSACTIONS OF SECTION ‘I, 673 
Consists essentially of a glass tube 50 cm. long, fixed vertically in a wide-mouthed 
glass vessel furnished with a second aperture and tubulure. The height of the 
vertical tube is invariable, but its width is regulated according to the amount of 
air required to be drawn through the apparatus in a given time. The bottom of 
this tube is closed with a platinum or silver plate pierced with a large number of 
very small holes, and two other similar perforated plates are inserted in the tube 
at certain intervals. The upper part of the tube is put in connection with an 
aspirating water-pump, and the absorbing liquid is placed in the lower glass 
yessel, whose second tubulure is connected with the supply of air in which the 
carbon dioxide has to be determined. When the aspirator is started the liquid is 
first drawn up into the vertical tube, and the air then follows through the perfo- 
rated plates which act as ‘scrubbers. Reiset, in his work, used baryta water as 
the absorbent, an aliquot part of which was titrated before and after the experi- 
ment, the changes in the volume of the liquid being corrected for by certait 
devices which I need not describe. 
The efficiency of the apparatus as a complete absorber of atmospheric carbon 
dioxide leaves nothing to be desired, but in dealing with large quantities of 
baryta solution, amounting to 400 c.c. or more, the errors due to inaccurate 
titrations, or to over or under estimation of the volume changes, are all thrown 
on the final result, of which they may form a considerable part. We have con- 
sequently altogether discarded the use of baryta as an absorbent in favour of 
caustic soda. The carbonate is estimated by a double titration process, suggested 
a few years ago by Hart, and we have succeeded in so far improving this method 
that there is no difficulty in determining in 100 c.c. of the alkaline solution an 
amount of carbonate corresponding to +4, c.c. of carbon dioxide. 
There is practically no limit to the amount of air which can be passed throug¢h 
an absorbing apparatus such as I have described, and one of very moderate 
dimensions will allow from 100 to 150 litres per hour to pass with perfect safety. 
Larger amounts can be dealt with either by increasing the size of the apparatus 
or by using several smaller ones arranged in parallel. 
With proper precautions, determinations can certainly be made to within 
‘02 part of carbon dioxide in 10,000 of air, so that with an apparatus of this 
kind it is possible to estimate the intake of carbon dioxide into a leaf or plant 
from ordinary atmospheric air, and to keep a sufficiently rapid stream of air 
passing over the leaf to maintain the tension of the carbon dioxide only slightly 
below the normal amount. 
The air is measured by carefully standardised meters, reading to about 20 c.c.; 
and since the amounts of air aspirated vary from 100 to 900 litres or more, there 
are practically no errors of measurement. The tension at which the air passes 
through the absorption apparatus is measured by a manometer, and all the 
volumes are reduced to standard temperature and pressure. 
All such experiments of course necessitate not only a determination of the 
carbon dioxide in the air which has passed over the leaf or plant, but also a 
simultaneous determination of the carbon dioxide in the ordinary air used. The 
accumulation of these air determinations clearly shows that the ordinary state- 
ments of our text-books as to the amount of carbon dioxide and its limits of 
variation are altogether misleading. 
In our experiments the air was in all cases taken from a height of four feet 
six inches from the ground, the amounts aspirated varying from 100 to 500 litres. 
In the month of July 1898 the minimum amount of carbon dioxide found was 
2-71 parts per 10,000 of air, and the maximum 2°86, During the winter months, 
when the ground was almost bare of vegetation, it rose to from 3:00 to 3'23 parts 
per 10,000; and on one foggy day, March 16, 1899, after a whole week of similar 
weather, we found the very exceptional amount of 3:62. Asa rule we may take 
it that the amount of carbon dioxide in the atmosphere during the period of 
greatest plant growth rarely falls short of 2:7 parts per 10,000, and rarely exceeds 
30 parts, with an average of about 2°85, These numbers come very close to the 
determinations of Reiset, and of Miintz and Aubin, and agree also fairly well with 
the Montsouris determinations, 
1899, seb 
