Adeney and Becker — Solution of Nitrogen and Oxygen. 389 



the capillary, and all traces of air expelled from the rounded end of the 

 rubber by repeatedly drawing water in and out of the capillary. 



The gas pipette, which is shown in position in fig. 1, was provided with a 

 conical tube, which allowed a gas-tight joint being made when pressed up 

 against the rubber-tipped capillary. The gases could then be extracted from 

 the water in the laboratory tlask and stored in the gas pipette by boiling the 

 water and drawing the gases over into the pump, and thence expelling them 

 into the pipette, by operating the mercury reservoir of the pump. Any excess 

 of water which distilled over was returned to the laboratory flask. "When 

 the gases were completely extracted and stored in the gas pipette, the latter 

 was removed and attached to the eudiometer, shown in fig. 2, by a similar 

 joint, and the gases drawn over and measured at reduced pressure. The 

 absorption pipettes were of similar construction to the gas pipette, and were 

 attached to the other capillary of the eudiometer. 



A number of tubes of 40 mms. internal diameter and of three different 

 lengths were used to expose the water to be tested. The lengths of the tubes 

 were 12 cms., 22 cms., and 32 cms., while the columns of water experimented 

 with were about 2 cms. shorter in each case. 



The water used was freed from air and from mineral and organic impurities 

 by distillation in vacuo, the tap water being boiled at about 30° C. in a 

 large glass flask, and the distillate collected in a glass receiver of about 

 4 litres capacity. In this way a large volume of purified water was 

 obtained, with an average air-content of less than 1 cc. of total gases 

 per litre. Each tube was filled by syphoning the water from the .bottle into 

 it, and allowing a fair quantity to overflow at the top. The tube was then 

 quickly closed by inserting a rubber stopper, provided with a glass stopcock, 

 without leaving any air-bubbles entrapped, and the tube was placed in a 

 thermostat until the contents attained the desired temperature, before 

 exposing the surface of the water to the air. 



If, as is commonly assumed, the surface layer of the water rapidly takes 

 up as much of the gas as is possible under the prevailing conditions, and the 

 dissolved gases only diffuse downwards through the mass of the water with 

 extreme slowness, the air-content of the water in the shallow tubes should 

 have been nearly as great as that of the water in the deep tubes, since the 

 lower layers should receive but little dissolved gas in the time during which 

 experiments were carried on. 



When a series of tubes were analysed, however, after exposing them to 

 the air for like periods of time, but too short to ensure complete saturation, it 

 was found that the air-content of the water in them varied approximately in 

 proportion to the depth of the water-column employed. Thus the tube con- 



