142 REPORT—1845. 
Report on the Gases evolved from Iron Furnaces, with reference to the 
Theory of the Smelting of Tron. By Prof. Bunsen, of Marburg, 
Hesse Cassel, and Dr. Lyon Puayratr, of the Museum of Giconomic 
Geology, department of Her Majesty’s Woods and Forests. 
In laying before the Association the report which we have now the honour 
to present, we are desirous, at the commencement of our subject, to examine 
closely the methods employed in the analyses of gases, not only as an argu- 
ment in favour of the processes used by ourselves, but also with the hope of 
improving the present state of eudiometry. 
Two distinct methods are employed in the analysis of combustible gases ; 
one of which consists in an exact determination of the voLumeEs of the gas 
about to be examined, and of those resulting from the combustion of its con- 
stituents with oxygen. By the other method, the products of combustion 
are collected in the liquid and solid form, and estimated directly according 
to WEIGHT. 
The last method would doubtless deserve the preference if we had to ope- 
rate upon a mixture of gases capable of being determined by the products of 
combustion without reference to the quantity of oxygen necessary to effect it ; 
in other words, when we have to examine a mixture containing only two 
combustible gases. In such a case, the combustion by means of oxide of 
copper affords products well-adapted for exact determination by weight. 
But, on the contrary, when the quantity of oxygen necessary for the com- 
bustion must be introduced as an element into the calculation, as is the case 
with the gases examined by us in the present paper, the method of analysis 
by weight is not only inexpedient, but also inexact. If that method were to 
be adopted, it is necessary to determine the loss (often not amounting to 
above a few centigrammes) sustained by a heavy combustion-tube, by weigh- 
ing it before and after the experiment, and thus subjecting it to all the sources 
of error due to a varying hygroscopic condition, and to the loss in weight oc- 
casioned by the long exposure of a considerable body of glass to a red heat. 
Another source of error equally great consists in the necessity for filling the 
whole apparatus for combustion and condensation with nitrogen gas previous 
to the commencement of the experiment. The smallest quantity of oxygen 
which may remain in the gas, or in the porous oxide of copper, or which 
may be introduced by diffusion, must derange the results, and cause great 
uncertainty in the determinations. Any error arising from this source is so 
much the more to be feared, because it does not affect one constituent merely, 
but extends its influence equally to the ascertained value of all the other in- 
gredients. 
We cannot afford better arguments for the reception of our methods of 
investigation than by briefly reviewing the results obtained by different in- 
quirers in the examination of the gases evolved from furnaces worked by 
charcoal. It is obvious that the composition of these gases cannot be the 
same under all circumstances, for the nature of the fuel, the pressure of the 
blast, and even the shape of the furnace itself, must exert a varying influence 
in modifying the processes which affect the composition of the gases. But 
when we consider, at the same time, that these modifying influences have 
their maximum and minimum in corresponding parts of furnaces treated in a 
similar manner, we still have a right to expect an elucidation of the law regu- 
lating the formation of the gases by a careful comparison of their compo- 
sition. One of us first endeavoured to solve this problem by an examination 
of the gases issuing from the furnace of Vickerhagen, although he did not 
then consider the results obtained in the inquiry as expressive of a general 
a 
