370 EXPLOSIVE GAS GENERATED— MACKAY. 



the enclosed air of this room vvliich is pretty pure. The lime 

 water is still transparent. The test is too crude, of course, to 

 show the existence of tlie small percentage of carbonic dioxide 

 in the air of the room. I breathe once into the mouth of the jar 

 and then shake up the lime water with the enclosed air. It has 

 instantly becDme quite milky. This lime water is, therefore, 

 a test cpood enousfh to indicate the existence of carbon diox- 

 ide if it should be present in a percentage as great as is 

 found in air that has Ijeen breathed aiul not largely diluted with 

 pure air. I therefore burn an inverted jar of the gas as in our 

 first experiment. If there is a hydro-carbon or carbon monoxide 

 present, they will be converted into carbon dioxide; and if the 

 proportion of carbon is at all considerable, its presence in the 

 heated and burnt air of the jar may possibly be demonstrated 

 The shaking up of the jar still leaves the lime water transparent, 

 which seems to suggest that the carbon compounds, if present, 

 cannot exist in any large quantities. 



5. Into this jar let us inject some pure nitrogen dioxide. As 

 the two transparent gases mix no change can be detected. But, 

 if the smallest bubble of air is injected the presence of oxygen is 

 indicated by the production of the characteristic fumes of nitro- 

 gfen tetroxide. Therefore we conclude that the gas contains no 

 oxygen when collected with due pi'ecautions. 



Our provisional conclusion from these rough experiments is, 

 that the gas is chiefly, if not altogether, hydrogen. 



The origin suggested is the decomposition of the water by the 

 over-heated metal in its circuit through the furnace, and the con- 

 version of metallic iron into its hydrated ferric oxide. If so, we 

 must expect that for every liter of gas produced, over one and 

 two thirds grams of iron are corroded ; that is, for every cubic 

 foot of gas nearly seven hundred grains of iron — 690 grains more 

 exactly. 



The conclusions finally suggested are, that conditions prevent- 

 ing the rapid formation of this gas (1) will lengthen the term of 

 service of furnaces ; (2) will obviate the danger of explosions 

 when air, owing to low water, may accumulate with the gas in 



