Pe ee eee eee 
a SE PN eR PE: ae NT ng SCE a I a ND meas Beal ae. fT a 
See nee aL et PP eae Se ne Se nal ene eee a Sm Serpe SNS iain a es ee eee meee Pe ps 
— 
Silliman and Wurtz on Air and Gas, . 45 
We have already alluded to the experiments of Mr. Carl 
Schultz, ae annex below a summary of his res 
M tz concludes from his experiments, that the loss 
_ within limits, the addition of air to very rich cannel gas, u 
] 
to 12 per cent of air, was followed by no loss of Motinateae 
power, but, onthe contrary, by a small gain. Thus boghead 
cannel gas, giving for 88 per cent gas an illuminating power of 
27-47 candles, or for 100 per cent gas 31°32 candles, with 12 per 
cent of air, gave 28:29 candles, showing that 12 per cent of at- 
mospherie air had increased the illuminating power of the flame 
by 0°82 candles. These results are obtained only by the use 
of an Argand burner. By substituting an intensity burner 
for the Argand, the results obtained with gas from boghead 
conform to the rule of half a candle loss for each one per cent 
of air. (Rev. Mr. Bowditch also states {generally, that “‘im- 
purities are far less destructive of light in Argand burners.”) 
Schultz sums up his results in the following propo- 
eeu 
. When coal gas is mixed with atmospheric air, its illu- 
atiathig power for all five feet burners is reduced in the pro- 
- portion of half a standard candle to every 1 per cent of air pres- 
ent, except in case of very rich gas burned with a 15 hole 
2. One cubic foot of atmospheric air will destroy an amount 
of light equal to 10 English standard candles, during one 
hour 
3, This loss: being constant, the percentage of aggregate 
~ loss will vary with the illuminating power of the gas used, 
While our results confirm in general those of Mr. Schultz in 
- 1860, and of Messrs, Audouin and Bérard, in 1862 ; it is ob- 
vious that the ratio of loss, with equal increments of air added 
to a 15 candle gas, is by no means constant. The difficulty of 
_ obtaining exact results, by the method of mixture of measured 
= 
volumes is so considerable, especially with quantities below 
4 10 per cent, that the only safe control of the results is that which 
= 
is obtained by eudiometrical analysis of the several mixtures test- 
ed, as was done by us in all cases both before and after the 
addition of air. It is obvious that the surprising loss of in- 
tensity by the addition to illuminating gas of small percent- 
ages of air must be owing not merely to the interior combus- 
tion due to the presence of oxygen, but still more, probably, to 
the associated nitrogen which acts not only asa dilu ent, or 
deductive quantity, but its specific heat is an actual divisory 
function in diminishing the flame temperature, The interest- 
