1915] on Problems of Hydrogen and the Rare Gases 559 



connects to an exhaust pump for the preliminary exhaust of the 

 space between D and the indiarubber. 



Everything being exhausted, the U-tube C is cooled in Hquid 

 air, while the ordinary air of the room is present in the gas-box 

 round the indiarubber. In half-a-minute a difference of pressure of 

 over 10 mm. is registered on the manometer. This, then, represents 

 the rate due to pure dry air diffusing in through the membrane and 

 through the cooled U-tube to the manometer. 



If C is not cooled, and A is simply surrounded with a beaker 

 lined with wet paper, then a pressure of 30-40 mm. in half-a-minute 

 is seen, the difference being due to the added rate of the aqueous 

 vapour Other vapours can be similarly tried by dropping a small 

 quantity of different liquids into beakers placed in turn round the 

 moanted membrane. Thus with ether, chloroform and ammonia 

 the manometer gives exceedingly rapid indications, about 5 cm. or 

 more being registered in a few seconds. Less active vapours are 

 benzene, alcohol, pyridine and acetonitrile. These are a few illus- 

 trations of what is no doubt a large field for future enquiry. A 

 viscous liquid like glycerin stops all diffusion when painted on 

 a dry film. The solution, however, in glycerin of substances such as 

 chloroform, etc., causes the manometer to act immediately. 



An interesting variation applicable to the subject of the discourse 

 is the passage over the membrane of respired air dried through a 

 tube in boiling solid carbonic acid (—80° C.). The 5 per cent or so 

 of carbonic acid therein contained gives an appreciable addition to 

 the rise of the manometer shown by dry air alone. If the breath be 

 '"held" for a half-minute or so J9efore expiring, the difference is 

 accentuated. When dry oxygen, hydrogen, and carbonic acid 

 respectively are passed througli the box over the membrane, the 

 corresponding pressures recorded are 10, 40, and over 100 mm. 

 respectively in the half -minute. Hydrogen therefore goes through, 

 roughly, at four times the rate of air, and carbonic acid at least ten 

 times as fast. 



More careful measures made with a McLeod gauge showed that 

 for the same membrane the rate expressed in c.c. per day through 

 each square centimetre at the ordinary temperature was 2*12 for air, 

 8 4 for hydrogen, and 30 • for carbonic acid. Different films give 

 different values within certain limits, according to their condition 

 and thickness, which last is of the order of O'Ol mm. The rate of 

 diffusion at different temperatures is readily determined by immersing 

 the gas-box containing the indiarubber membrane in a fluid the 

 temperature of which can be properly regulated. The variation of 

 rate with temperature is found to be logarithmic, with a definite 

 break in the curve at 0° C, and also at other points below this in the 

 case of hydrogen and carbonic oxide ; as far as experiments have 

 been carried out. Helium has a slightly greater rate than air, while 

 carbonic oxide and nitrogen are not far removed. Some of these 



