Transmission of Gases through Water. 361 



streaming effects through the water. He explains these as 

 being cansed by the water becoming heavier as it dissolves the 

 gas. and sinking to the bottom. 



Hiifher, in order to avoid this streaming effect, employed a 

 thin plate of the porous mineral hydrophane to hold up a 

 column of water in an experimental tube, and so provided the 

 means oi introducing the gas at the bottom of the water and 

 of studying the upward diffusion of the gas. 



The author lias met with this downward streaming in the 

 course of some experiments recently made upon the solution 

 of atmospheric gases by water. It did not, however, appear to 

 him that the cause of the streaming, as suggested by Hiifner, 

 offered a wholly satisfactory explanation of the phenomenon 

 observed. He accordingly thought that a careful investi- 

 gation of this streaming effect might be followed by results 

 of some value. 



Method of Experiment. 



Glass tubes about 2 metres long and 18 millions, bore 

 have been employed for all the experiments to be described, 

 except for a few preliminary ones. The dimensions of these 

 will be given with the descriptions of the experiments. 



The tubes were nearly filled either with distilled water or 

 with sea-water, and warmed and exhausted as far as desirable 

 by means of a mercury-pump. The corks closing the lower 

 ends of the tubes were fitted with tine glass tubes to provide 

 the means of attachment to the mercury-pump for exhaustion, 

 and for drawing off samples. 



After the gases in solution and in the air-spaces had been 

 abstracted, the corks closing the upper ends of the tubes were 

 removed, and the tubes for most of the experiments were 

 immersed to a depth of about 1300 mm. in a large water- 

 jacket, which was kept cool by a constant flow through it of 

 water from the main supply. 



The object of the water-jacket was to prevent the for- 

 mation of downward convection-currents, by keeping the 

 lower portions of the water-columns in the tubes at a slightly 

 lower temperature than the upper portions, which were 

 exposed to the temperature of the laboratory. 



Jn most of the experiments to be described, the surfaces of 

 the water-columns were kept constantly agitated by means 

 of currents of air drawn through their upper layers. For the 

 Few preliminary experiments to be described, the tubes were 

 fitted with corks carrying fine glass inlet and exit tubes 

 of '.) mm. 1 ore, and for the later experiments with glass 

 capillary tubes of 0*5 mm. bore. The inlet-tubes have 



