132 
THE CAUSE OF THE LUMINOSITY OF FLAME. 
flame. We go on working the air-pump, reducing the pressure still further until we 
get it down to a quarter of an atmosphere, and you see that the luminosity of the flame 
suffers a further great reduction. Thus, while we had a very brilliant flame when 
arseuic was burning in oxygen at the ordinary pressure, we have steadily reduced it to 
its present state of enfeebledness by rarifving the oxygen, and thus expanding the 
vapours of the arsenic and arsenious acid. We shall presently see the bearing of this 
experiment upon the luminosity of the flame of coal-gas. [In this striking experiment 
the arsenic was burned in a bulb tube, connected with a cylindrical receiver, from which 
the vapour produced was withdrawn by the aid of the air-pump. The supply of oxygen 
■was so regulated by a stop-cock that the arsenic flame preserved the same volume 
throughout the experiment.] I will now reverse this experiment, and show you how 
we can gradually increase the lumiuosity of a flame by increasing the density of the 
vapour composing it. I place this small spirit-lamp, which gives us at present such a 
trifling amount of light, beneath the receiver of a condensing air-pump, and you notice 
that the flame becomes brighter as the pressure of the surrounding air is increased. 
[The experiment was performed as described, the result being that the feeble spirit- 
flame became very bright.] 
“ Let us now inquire whether this relation between density and luminosity subsists in 
gas flame. Coal-gas is essentially a mixture of certain hydrocarbons or compounds of 
carbon and hydrogen, the flames of which, together with their chemical formulae, will 
be found in our table. [See next page.] 
“ Now, it is well known to chemists that one hydrocarbon may be transformed into 
another of greater complexity, and consequently of greater vapour density, by direct 
condensation, and that this condensation is effected by heat. Thus, marsh-gas, the 
hydrocarbon which comes first on our list, when raised to a very high temperature, 
is converted into acetylene, a gas which contains twice the quantity of carbon in the 
same standard volume. Under the influence of heat the hydrogen gradually separates 
from the carbon, while the atoms of the latter element shrink, as it were, more closely 
together, so that two volumes of marsh-gas become condensed into one volume of acety¬ 
lene. The condensation of marsh-gas may, however, be carried much further. Thus 
w r e may get the hydrocarbon called naphthalene, the vapour of which contains ten times 
as much carbon in a given space as marsh-gas contains. Taking the density of hydro¬ 
gen as unity, we find that the relative densities of marsh-gas, acetylene, and naphthalene 
are 8, 13, and 64 respectively. The heterogeneous body called pitch produced in the 
manufacture of coal-gas contains the hydrocarbons chrysene and pyrene, in which the 
carbon atoms are more closely packed than in naphthalene. They are perfectly vola¬ 
tile. and their vapours necessarily very dense. To the presence of such compounds I 
am inclined to ascribe the luminosity of the gas flame. I believe that the lighter hy¬ 
drocarbons, when exposed to the intense heat of the flame, are transformed into dense 
vapours which have the luminiferous character of arsenic-vapour or phosphorus-vapour. 
But, you may say, how is it that we get a sooty deposit on a piece of white porcelain if 
there be no suspended particles of carbon in the flame ? There is no reason to suppose 
that soot or lamp-black is free carbon. It invariably contains hydrogen, and if we wish 
to convert it into pure carbon we must subject it to prolonged ignition. Even this 
treatment fails to remove all traces of hydrogen, and to obtain chemically pure carbon 
the ignited product must be simultaneously subjected to the action of chlorine. So 
that the substance which is commonly regarded as uncombined carbon is doubtless an 
agglomeration of various heavy hydrocarbons formed in the interior of the flame. 
When the flame is allowed to play upon a porcelain plate or other cold body, the 
vapours are condensed into soot, and when the flame is cut in two by wire-gauze they 
are condensed into a thick smoke. In the crude product which arises from the retorts of 
the gas-works we may observe the same kind of smoke, which is, in fact, a fog of heavy 
hydrocarbons. 
“ The popular belief in the dependence of luminosity upon incandescent solid par¬ 
ticles seems, moreover, inconsistent with certain obvious facts. Taking the gas flame 
from a common fish-tail burner, we find it so perfectly transparent that the smallest 
print can be read through any part of it without the least difficulty. We also find 
that the amount of light received by a body placed opposite to the flat side of the fish¬ 
tail flame is not diminished when the edge of the flame is turned towards the body. 
Now, if the light proceeded from incandescent solid particles, the most luminous part of 
