which are evolved in Coal- Mines. 141 



From these experiments, it results that carbonic oxide gas, 

 mixed with carburetted hydrogen gas, cannot be determined 

 quantitatively by means of potassium ; that it is only indicat- 

 ed by a diminution of the volume of gas when it is present in 

 considerable quantity ; and that that diminution amounts to 

 so much the less, the more strongly and the more uninterrupt- 

 edly the potassium is heated ; nay, that in this case an in- 

 crease of volume may take place even when the carbonic oxide 

 gas amounts to 14 per cent. 



What, then, can be the cause of this increase ? Moisture, 

 which in the first experiment (§ v.) seems to have had influ- 

 ence, was entirely counteracted in the subsequent experiments 

 by boiling the mercury, and by chloride of calcium. There can, 

 therefore, only have taken place a partial decomposition of 

 the carburetted hydrogen gas by means of the heated potassium ; 

 for if we heat a mixture of pit-gas from the Wellesweiler Siollen 

 and carbonic oxide ga6 to the same point, without the presence 

 of potassium, no increase of volume presents itself. It appears 

 that here the potassium acts in the same manner as do metals 

 on ammoniacal gas when heated in it. In the experiments in 

 which potassium was heated in one of the two pit-gases with- 

 out the addition of carbonic oxide gas, there was always pre- 

 sented a blackening of the tube in the neighbourhood of the 

 metal. 



As, therefore, potassium is not adapted for the separation 

 of carbonic oxide gas from an inflammable gaseous mixture, 

 there remain no other means for the analysis of any combina- 

 tion of the four inflammable gases, hydrogen, carbonic oxide, 

 carburetted hydrogen, and olefiant gas, except chlorine, deto- 

 nation with oxygen, and the determination of the specific 

 gravity. 



I have instituted experiments on the pit-gas of Gerhard's 

 Siollen at a red heat, similar to those made with the previous 

 gas. 



The dried gas was conducted 10 times in one experiment, 

 and 8 times in another, through the red-hot porcelain tuhe ; 

 but after the fourth transmission the volume was no longer 

 increased. This increase was somewhat smaller in the second 

 experiment than in the first, which was probably caused by 



