GAS, COAL 569 



this solution, a small tost tube is filled with quicksilver, and inverted in the mercury 

 trough ; a few drops of a saturated solution of pyrogallic acid in water are thrown up 

 into this tube by means of a pipette, and then a similar quantity of a strong solution 

 of potash ; a coke bullet attached to a platinum wire is introduced into this liquid, 

 and allowed to saturate itself ; it is then withdrawn, and conveyed carefully below 

 the surface of the mercury into the eudiometer containing the residual gas of experi- 

 ment No. 1 ; every trace of oxygen will be absorbed in a few minutes, when the 

 bullet must be removed, and the volume being again measured, the diminution from 

 the last reading will represent the amount of oxygen originally present in the gas. 

 It is essential that the coke bullet, after saturation with the alkaline solution of 

 pyrogallic acid, should not come in contact with the air before its introduction into 



III. Estimation of the Luminiferous Constituents. 



Various methods have been employed for the estimation of the so-called olefiant 

 gas (luminiferous constituents) contained in coal-gas. The one which has been 

 most generally employed, depends upon the property which is possessed by olefiant 

 gas, and most hydrocarbons, of combining with chlorine, and condensing to an oily 

 liquid : hydrogen and light carburetted hydrogen are both acted upon in a similar 

 manner when a ray even of diffused light is allowed to have access to the mixture ; 

 but the condensation of the olefiant gas and hydrocarbons takes place in perfect 

 darkness, and advantage is therefore taken of this circumstance to observe the amount 

 of condensation which takes place when the mixture is excluded from light. The 

 volume, which disappears during this action of the chlorine, is regarded as indicating 

 the quantity of olefiant gas present in the mixture. There are many sources of error 

 inseparably connected with this method of operating, which render the results 

 unworthy of the slightest confidence ; the same remark applies also to the employment 

 of bromine in the place of chlorine ; in addition to the circumstance that these 

 determinations must be made over water, which allows a constant diffusion of 

 atmospheric air into the gas, and vice versa, there is also formed in each case a 

 volatile liquid, the tension of the vapour of which increases the volume of the residual 

 gas ; and this increase admits of neither calculation nor determination. The only 

 material by which the estimation of the luminiferous constituents can be accurately 

 effected is anhydrous sulphuric acid, which immediately condenses the luminiferous 

 constituents of coal-gas, but has no action upon the other ingredients, even when 

 exposed to sunlight. The estimation is conducted as follows : A coke bullet, 

 prepared as described above, and attached to a platinum wire, being rendered 

 thoroughly dry by slightly heating it for a few minutes, is quickly immersed in a 

 saturated solution of anhydrous sulphuric acid, in Nordhausen sulphuric acid, and 

 allowed to remain in the liquid for one minute ; it is then withdrawn, leaving as little 

 superfluous acid adhering to it as possible, quickly plunged beneath the quicksilver 

 in the trough, and introduced into the same portion of dry gas, from which the 

 carbonic acid and oxygen have been withdrawn by experiments I. and II. ; here 

 it is allowed to remain for about two hours, in order to ensure the complete absorption 

 of every trace of hydrocarbons. The residual. volume of gas cannot, however, yet be 

 determined, owing to the presence of some sulphurous acid derived from the decom- 

 position of a portion of the sulphuric acid : this is absorbed in a few minutes by the 

 introduction of a moist bullet of peroxide of manganese, which is readily made by 

 converting powdered peroxide of manganese into a stiff paste with water, rolling it 

 into the shape of a small bullet, and then inserting a bent platinum wire, in such a 

 manner as to prevent its being readily drawn out ; the ball should then be put in a 

 warm place, and allowed slowly to dry, it will then become hard, and possess con- 

 siderable cohesion, even after being moistened with a drop of water, previous to its 

 introduction into the gas. After half an hour, the bullet of peroxide of manganese 

 may be withdrawn, and replaced by one of caxistic potash, to remove the watery 

 vapour introduced with the previous one ; at the end of another half-hour, this bullet 

 may be removed, and the volume of gas at once read off. The difference between 

 this and the previous reading, gives the volume of the luminiferous constituents 

 contained in the gas. This method is very accurate ; in two analyses of the same 

 gas, the percentage of luminiferous constituents seldom varies more than O f l or O f 2 

 per cent. 



IV. Estimation of the Non-Luminiferous Constituents. 



These are light carburetted hydrogen, hydrogen, carbonic oxide, and nitrogen. 

 The percentages of these gases are ascertained in a graduated eudiometer, aboiit 2 

 feet in length, and ^ths of an inch internal diameter ; the thickness of the 

 glass being not more than ith of an inch. This eudiometer is furnished as 



