Dialytic Separation of Gases by Colloid Septa. 529 



wire (No. 21), measuring 4*1 cub. centiois., the iron was heated 

 in a small glass tube, to exclude the idea of the conceivable per- 

 meability of the porcelain tube. The iron gave off gas at a 

 pretty uniform rate, which amounted in an hour to 29 '8 cub. 

 centims., of which 4*44 cub. centims. were carbonic acid, and 

 the remainder principally carbonic oxide, with hydrogen and a 

 trace of a hydro- carburet. Here the iron wire gave off 7-27 

 volumes of gas. 



3. In a third experiment, on thin iron wire (No. 23), the ex- 

 trication of the natural gases at a red heat was pushed to a greater 

 degree of exhaustion. The weight of the iron was 39 grms., and 

 its volume 5 cub. centims. In the first and second hour the gas 

 collected was 45 cub. centims.; in the third hour 10 - 85 cub. 

 centims. ; in the fourth and fifth hours 5*65 cub. centims. • in 

 the sixth hour 0*9 cub. centirn., and in the seventh hour 07 cub. 

 centim. The iron appears to be now nearly exhausted, after the 

 extrication of 63*1 cub. centims., or 12*55 volumes of gas. 



It is evident that iron cannot be safely dealt with in experi- 

 ments upon the permeation or upon the absorption of gases, till 

 these gases, whether self-produced or preexisting, are first extri- 

 cated from the metal. The carbonic oxide observed in the tube 

 experiments of M.Deville may have been derived from the same 

 source*. 



4. To observe the absorption of hydrogen, the mass of ex- 

 hausted iron wire remaining after the last experiment was heated 

 to redness, and cooled gradually in the same gas. The metal 

 was afterwards freely exposed to air (as usual) to get rid of any 

 loosely attached hydrogen. Now exhausted again by the Spren- 

 gel pump at a low red heat, the iron gave 2*5 cub. centims. of 

 gas in one hour, but the greater portion in the first ten minutes, 

 consisting of 



2*3 cub. centims. Hydrogen. 



0*2 „ Carbonic oxide &c. 



2-5 



The iron appears therefore to be capable of holding 0*46 vol- 

 ume of hydrogen. The wire became white, like galvanized iron. 

 This was confirmed in a second observation, a thicker wire hold- 

 ing 0*42 vol. hydrogen. 



5. The same specimen of iron was now charged with carbonic 

 oxide gas, in the manner it had previously been charged with 

 hydrogen. It was also freely exposed to air. The iron wire re- 

 mained soft, was not capable of becoming hard when heated red- 



* The gases which escape from cast iron in a state of fusion have been 

 examined by M. L. Cailletet. They appear to contain from 49 to 58 per 

 cent, of carbonic oxide, 34 to 39 of hydrogen, and 8 to 12 of nitrogen. — 

 Comptes Rendus, vol. lxi. p. 850 (1865). 



Phil. Mag. S. 4. No. 219. Suppl. Vol. 32. 2 M 



