ge 
ia i a ail at 
; 
ON THE GASES EVOLVED FROM IRON FURNACES. 159 
sometimes observed in the succeeding eudiometric analyses. The gases pro- 
cured after this treatment, consisting of hydrogen, light carburetted hydro- 
gen and carbonic oxide, are entirely destitute of smell, and without action 
upon mercury. As soon as all the atmospheric air is expelled from the ap- 
paratus, which we find by analysis to be effected by the distillation of about 
300 grains of coal, the conducting tube () is dipped under mercury and the 
gas collected. In order to have it of average composition, the gas is col- 
lected over mercury in a glass vessel, of a capacity of 800 to 1000 cubic cen- 
timetres. 
The glass tubes conveying the gas into the vessel is connected with the rest 
of the apparatus by means of a caoutchouc joint, and a tube, rather narrowed 
in the middle. This contracted tube is fused when the receiver is filled, 
but immediately opened again with a pair of tongs in that part which still 
remains in contact with the system of absorption, so that the experiment may 
be continued until the coal ceases to yield gas. As soon as this point is at- 
tained, the fire is removed from the combustion furnace, and the distillatory 
tube opened by cutting away with a diamond its drawn-out neck, so far as it 
is filled with coal-tar. The part of the absorptive system formerly in con- 
nection with the mercurial apparatus is now attached to a hand air-pump, 
and the apparatus filled with atmospheric air by a few gentle strokes of the 
pump. The loss in weight of the distillatory tube, after being filled with air, 
adding the weight of the part cut off, gives the amount of coal left behind 
by the distillation, and also the total weight of the liquids and gases which 
have escaped from the coal. The quantity of fluid matter is determined by 
the weight of the receivers 6, c, and by the loss in weight of the fragment of 
glass tube when freed from tar. The receivers d, e give the quantity of car- 
bonic acid and sulphuretted hydrogen, the receiver f,g, 4 the weight of the 
olefiant gases and condensable hydrocarbons. By subtracting the weight of 
these collected products of distillation from the loss sustained by the distil- 
latory tube, the remainder indicates the weight of the non-condensable gases, 
the composition of which in hydrogen, carburetted hydrogen and carbonic 
oxide, is easily determined by a eudiometric analysis. 
The amount of tar produced by the distillation may be determined by 
throwing the contents of the first receiver on a weighed filter moistened with 
water, washing it, and, after drying both it and the moist receiver, the weight 
of these, added to that of the tar in the cut fragment of tube, gives a very 
exact result as to its amount. The ammonia contained in the water is 
best obtained by distilling it with a large excess of potash into a receiver 
containing muriatic acid, until at least two-thirds of the liquid have passed 
over, and it is then collected in the usual way by evaporation and precipi- 
tation with chloride of platinum, the washing of the double salt being best 
effected by a mixture of alcohol and ether, according to Varrentrapp and 
Wills’ recommendation. The amount of water is of course known by de- 
ducting the weight of the tar and ammonia from the total weight. 
In order to draw conclusions as to the composition of the gases of the fur- 
nace, it is of importance to ascertain the composition of those absorbed by 
the perchloride of antimony. To determine this point, a quantity of coal 
was heated to redness with the precautions already described, and collected 
in a gasometer filled with milk of lime. This gas, carefully dried by passing 
over chloride of calcium, was led into perchloride of antimony until the 
latter was saturated. An indefinite quantity of the black liquid thus ob- 
tained was put into a combustion-tube with oxide of copper, the front part of 
the tube being supplied with copper shavings, and on combustion, 0°1226 
_ water and 0°3626 carbonic acid were obtained, which correspond to 
