Dialytic Separation of Gases by Colloid Septa. 419 



therm. 20°. For a square metre the rate is 2*22 and 2*29, ave- 

 rage 2*25 cub. centims. per minute. The proportion of oxygen 

 was, in the first experiment 42*5, and in the second 41*66 per 

 cent. 



When the same dialyzing-bag was kept at a temperature of 

 60° C, the volume of air collected in seven minutes was 6*22 

 and 7*06 cub. centims. For the square metre this amounts to 

 6*21 and 7*05, mean 6*65 cub. centims. per minute. The pas- 

 sage of air through rubber is therefore almost exactly three times 

 as quick at 60° as at 20° C. 



Again, the dialyzing-bag was kept at 4° C. by being sur- 

 rounded by ice and salt. The air now collected in seventy-two 

 minutes was 5*78 and 5*77 cub. centims. in volume — for a square 

 metre, 056 cub. centim. per minute. The passage of air through 

 rubber thus appears to be four times as slow at 4°C. as it is at 20°. 

 The proportion of oxygen in the dialyzed air increased at the same 

 time. In the two portions of air collected at 4° the oxygen was 

 46" 75 and 47*43 per cent. The increase of oxygen at a low tem- 

 perature was confirmed in other experiments ; but it appeared at 

 the same time that the rubber was liable to acquire a true poro- 

 sity to a slight extent when retained for some hours about 0° C. 

 The rubber then allowed air to pass through it containing no 

 more than 28 or even 23 per cent, of oxygen, and in volume still 

 very small. The rubber has become rigid by the cold, and is 

 now acting feebly as a porous substance, allowing a little gas- 

 diffusion to take place through its substance. Such a condition, 

 which is accidental to caoutchouc at a low temperature, appears 

 to be constant with gutta percha, a harder material, at 20° C, 

 and even higher temperatures. 



A large bag of varnished silk with a surface of 1*672 square 

 metre (two square yards) was found still more convenient. It 

 was, however, rather beyond the exhausting-power of the largest 

 Sprengel-pump. It yielded in eight minutes, without any col- 

 lecting-flask, 22, 21*55, and 21*5, mean 21*68 cub. centims. 

 This was a supply of 2*71 cub. centims. per minute, and was at 

 the rate, for a square metre, of 1*62 cub. centim. per minute. 

 The supply would have been about a half more if the dialyzed 

 air had not gained upon the pump. The air of the first and last 

 observations contained respectively 41*89 and 41*85 per cent, of 

 oxygen. 



The usual proportion of oxygen in air dialyzed by rubber ap- 

 pears to be about 41*6 per cent. ; and it may be described as 

 atmospheric air deprived of one-half of its usual proportion of 

 nitrogen. A single dialysis of air therefore carries the experi- 

 menter already halfway from air to pure oxygen as the final 

 result. But the gain by a second dialysis could not be so great, 



2E2 



