Dialytic Separation of Gases by Colloid Septa. 505 



gas admitted into the annular space between the porcelain and 

 platinum tubes was either atmospheric air or hydrogen at the 

 natural temperature. The tubes being placed across an empty 

 furnace, the latter was now lighted ; and it was seen that, with 

 air circulating outside the platinum, the vacuum remained undis- 

 turbed, even when the temperature of the tubes rose to a bright 

 red heat. But when dry hydrogen was driven through the same 

 annular space, the platinum, while continuing impermeable at all 

 temperatures below a dull red heat, began to admit hydrogen to 

 the vacuum as soon as the external porcelain tube became visibly 

 red-hot. In seven minutes the Sprengel pump now delivered 

 15-47 cub. centims. of gas, of which 15*27 cub. centims. ap- 

 peared, by explosion with oxygen, to be hydrogen. 



In a repetition of the last experiment, hydrogen dried by sul- 

 phuric acid was again allowed to circulate in excess outside the 

 platinum. After a vacuum was once obtained within the plati- 

 num tube, the gas delivered by the Sprengel pump, in the cold, 

 during a period of forty minutes, amounted to no more than a 

 bubble of the size of a pin-head, showing the tightness of the 

 apparatus. The Sprengel pump being constantly kept in action, 

 the tubes were now heated to redness, and then gradually to a 

 temperature approaching a white heat. The gas delivered each 

 five minutes was found to be 13, 15*5, 17*4, 169, 18'6 cub. 

 centims. as the temperature rose. These volumes are referred to 

 a temperature of 20° and barometer of 760 millims. The last 

 observation gives a passage of 372 cub. centims. of hydrogen 

 per minute. The platinum tube employed here was joined with- 

 out solder, having been drawn from a mass of platinum which had 

 been aggregated by fusion. It was similar in this respect to the 

 tube employed by M. Deville. The tube was 0-812 metre in 

 length (32 inches) and 1*1 millim. in thickness, with an internal 

 diameter of 12 millims. But only a portion of about 200 mil- 

 lims. (8 inches) of the tube was heated to redness in the fur- 

 nace experiment. The inner surface of the heated portion had 

 therefore an area of 00076 square metre. Hence one square 

 metre of heated platinum delivered 489*2 cub. centims. of hydro- 

 gen per minute. This result admits of comparison with the 

 passage of gases through a septum of rubber. In the most fa- 

 vourable circumstances, when the thin membrane of a rubber 

 balloon was employed, the passage of air into a vacuum was at 

 the rate of 26 - 5 cub. centims. per square metre in one minute. 

 The passage of hydrogen may be taken as 4*8 times as rapid as 

 that of atmospheric air, or at 127*2 cub. centims. per minute. 

 But while the thickness of the platinum septum was 1*1 millim., 

 that of the rubber film was only one- seventieth part of a milli- 

 metre. Hence we have the ultimate comparison : — 



