Chemical Examination of the Fire-Damp. 209 



such very slight circumstances, that a small matter will cause a 

 ball to be wholly inert which would otherwise have acted with 

 effect. 



In applying nitrous gas to determine the quantity of oxygen in 

 fire-damp, I employed the method of Dr. Dalton, as described by 

 Dr. Henry's Elements of Chemistry, A measured quantity of 

 fire-damp was added to the nitrous gas contained in a graduated 

 tube half an inch wide, and the gases were allowed to act on 

 each other over water, without agitation. The diminution of 

 volume had attained its maximum in five or six minutes, and in 

 general much sooner. Of the total loss, i|th were taken as oxy- 

 gen. This method is not in all cases rigidly correct, but its in- 

 dications were sufficiently exact for my purpose, controlled as 

 they were by the action of platinum, by the analysis of the, gas 

 by detonation with oxygen, and by the specific gravity of the 

 gases. Before relying at all on this method, however, I applied 

 it in the analysis of gaseous mixtures containing known quanti- 

 ties of oxygen gas. On applying it to the analysis of atmospheric 

 air it indicated 20-4 per cent, of oxygen. On agitating the air 

 and nitrous gas, just after admitting them into the same tube, the 

 diminution in volume was excessive. In a specimen of nitrogen 

 gas, to which so much air was admitted that the whole mixture 

 contained 3 per cent, of oxygen, nitrous gas indicated 3-3 per cent, 

 of oxygen in one experiment, and 3-2 in a second. With nitro- 

 gen, which contained 3-6 per cent, of oxygen, nitrous gas indica- 

 ted 4-4 in one trial, and in a second 4-1 per cent, of oxygen. In 

 nitrogen gas, with 4-T per cent, of oxygen, nitrous gas indicated 

 4.7 per cent, in one trial, and 5*2 in the second. In nitrogen 

 containing 7*3 per cent, of oxygen gas, nitrous gas indicated 

 7*4 in the first experiment, and 8-4 in the second. In the last 

 case a large excess of nitrous gas was employed. In nitrogen gas 

 in one experiment, and 11-5 in a second.* In this last case also 

 nitrous gas was used in large excess. 



In these experiments the error is very uniformly such, that 

 more oxygen was indicated than was actually present. The 

 causes of error appear to be especially twofold, — agitation, and a 

 large excess of nitrous gas. By permitting the action to ensue 

 tranquilly, and avoiding much excess from nitrous gas, the indi- 



* There appears to be some omission here. — Edit. Phil. Mag. 

 Vol. XXXVII, No. 2.— July-October, 1839. 27 



