566 Thoday . — ( 9/2 the Capillary Eudiometric Apparatus of 
his experiments to be the initial composition of the air. Lamarliere, 1 
working in Bonnier’s laboratory, found 0-04 % C 0 2 and 20-60 % 0 2 . 
Puriewitsch, using Baranetzky’s modification of the apparatus, 2 determined 
the initial composition of the air in all his experiments, and in these 
analyses the percentage of oxygen is in no case higher than 20-6. More- 
over, Stich 3 says that numerous analyses of atmospheric air (with an older 
form of the apparatus) differed from Dumas’s by 02-0*3 %. 
Some time ago I required such a method of analysis for work on 
respiration, and Dr. F. F. Blackman suggested this apparatus as possibly 
a suitable and convenient one if it would give sufficiently accurate results. 
Earlier trials with it in the Cambridge Botany School had not been 
altogether successful. 4 My experience was also unfavourable at first. The 
directions given by Aubert were closely followed, but a number of difficulties 
were encountered. These were only overcome by modifying the procedure. 
Recently, however, I have succeeded in obtaining results, the degree of 
accuracy of which approaches very nearly to the highest to be expected ; 
I hope, therefore, that this account of the technique adopted may prove 
of service. 
Fig. 1 illustrates the essential features of the apparatus. By turning 
the handle A, the piston B is moved in the cylinder C, which is filled with 
mercury. In this way an air-thread may be propelled from the graduated 
part, EF, of the capillary tube into the bent ungraduated part, FG, where 
absorption takes place, or drawn back into the graduated part again for 
measurement. The end of the capillary tube opens under mercury in the 
reservoir H. 
The measuring and absorbing regions of the capillary tube correspond 
to the distinct measuring and absorbing tubes of the earlier forms of the 
apparatus. The absorbing liquid is drawn into the capillary, from a test- 
tube which is pressed down over the end of the capillary in the reservoir H, 
and is then expelled, the air being brought into contact with the film of 
liquid left wetting the walls of the absorbing region. The bulb D acts as 
a trap for the gas under analysis to prevent it being inadvertently drawn 
down into the metal cylinder and there partially lost. It is not to be used 
for absorption, as stated by Macdougal. 6 The absorbing solutions should 
never reach the graduated part of the tube, which is used solely for 
measurement. 6 
1 Rev. g&i. de Bot., iv, 1892, p. 481. 2 Jahrb. f. wiss. Bot., xxxv, 1900, p. 578. 
3 Flora, N.S., xlix, 1891, p. 7. 
4 Cf. Darwin and Acton : Physiology of Plants, 1901, p. 10. 
5 Textbook of Plant Physiology, 1901, p. 259. 
6 Stoward (Ann. of Bot., 1908) found a constant error of — o-86 per cent, in the amount of C0 2 , 
and attributes this to potash persistently retained in the bulb ! Unless Stoward followed Macdougal’s 
directions it is difficult to understand either how potash got into the bulb at all, or how such a large 
error could have resulted if the tube and bulb were thoroughly washed out with acid. 
