168 Experiments on the Gas from Coal. 



to avoid errors from its absorption in rising through the water. 

 To this is admitted half a cubic inch (equivalent to 50 measures) 

 of the gas under examination, and the mixture is left, excluded 

 from the direct light of the sun and perfectly quiescent, for fifteen 

 minutes. At the expiration of this time, the remainder is noted, 

 and the diminution which has taken place being divided by 2, 

 the quotient shows the quantity of defiant gas in fifty measures 

 of the mixture. This ])rocess, I am cavvare, however, does not 

 give results of perfect accuracy; for, in addition to other sources 

 of fallacy, I find that chlorine begins to act on sulphuretted hy- 

 drogen much sooner than is generally supposed*, though within 

 the period mentioned, and in such narrow tubes, it does not oc- 

 casion a sensible diminution of bulk. The method described 

 may, therefore, be considered as affording a tolerably near ap- 

 proximation to the proportion of olefiant gas ; and as all the va- 

 rieties of coal gas were subjected to the test under precisely the 

 same circumstances, the errors must have been of nearly the 

 same amount in all cases, and cannot materially interfere with 

 the fair comparison of the different specimens of coal gas, so far 

 as respects their proportion of olefiant gas. 



3. To ascertain the quantity of comhustihle matter in gas 

 which had been deprived only of sulphuretted hydrogen and car- 

 bonic acid, a mixture of the gas with a due proportion of oxygen 

 gas was fired by the electric spark over mercury. This method 

 I preferred to slow combustion, carried on with the apparatus 

 which I have described in the Philosophical Transactions for 

 180S, solely because, when a great number of experiments are 

 necessary, as in this inquiry, the method of detonation is attended 

 with a great saving of time. But on all occasions where only 

 few experiments are required on gases of great combustibihty, I 

 prefer slow combustion, both on account of greater safety to the 

 apparatus, and, from the quantities that may be consumed, of 

 greater accuracy also. When rapid combustion is practised, I 

 believe that, on the whole, more accurate results are gained bv 

 firing the gas at one operation properly conducted, than at two. 

 The latter method seems to have been preferred by M. Berthol- 

 let; but so far as my experience goes, it is more apt to precipi- 

 tate charcoal from the gas. 



To burn each measure of the early and more combustible pro- 

 ducts of gas, I employed from 3 to 4 measures or upwards of 

 oxygen gas the degree of purity of which had been ascertained. 

 The volume being noted after firing, and again after agitating 



* While this sheet was passing through the press, I have noticed a pas- 

 sage in Mr. Brande's Manual of Chemistry (page 156 n.), fioin which it ap- 

 pears that the speedy action of chlorine on carburetted hydrogen had been 

 observed by Mr. Faraday. 



the 



