22 MR SWAN ON THE CONSTITUTION OF FLAME. 
flame, founded on the principle, that ‘“‘ There is a connexion between the re- 
frangibility of the light which a burning body yields, and the intensity of the 
chemical action going on; and that the refrangibility always increases as the 
chemical action increases.’”’* On this principle, Professor Draper reasons as fol- 
lows :—‘“ All common flames, as is well known, consist of a thin shell of ignited 
matter, the interior being dark, the combustion taking effect on those points only 
which are in contact with the air. From the circumstances under which the air 
is usually supplied, this ignited shell cannot be a mere mathematical superficies, 
but must have a sensible thickness. If we imagine it to consist of a series of 
strata, it is obvious that the phenomena of combustion are different for each. 
The outer stratum is in absolute contact with the air, and there the combustion is 
most perfect ; but by reason of the rapid diffusion of gases into one another, cur- 
rents, and other such causes, the atmospheric air must necessarily pervade the burn- 
ing shell to a considerable depth; and in the successive strata, as we advance in- 
wards, the activity of the burning must decline. On the exterior stratum oxy- 
gen is in excess, at the interior the combustible vapour, and between these limits 
there must be an admixture of the two, which differs at different depths.” Ad- 
mitting the principle already enunciated, Professor Drarer adds,— It follows, 
that each point of the superficies of every flame, no matter what the combustible 
may be, must yield all the colours of the spectrum, the violet coming from the 
outer strata, the yellow from the intermediate, and the red from those within. 
If we could isolate an elementary horizontal section of a flame, it should exhibit 
the appearance of a rainbow ring.” + 
The results [ had obtained in my experiments, led me to doubt the correct- 
ness of these views; for I found, that while the envelope of a hydro-carbon flame 
formed an absolutely continuous spectrum, extending from about the line C nearly 
to the line H of Fraunnorer, the bright inner cone formed a totally independent 
spectrum of superior brilliancy, also quite continuous, and of still greater extent; 
and the blue cone had a remarkable interrupted spectrum, peculiarly its own, whose 
extreme visible portions lay, respectively, between the lines D and E, and close to 
the line G of FRauNHoFER. While my experiments thus amply confirmed what 
was already known regarding the very diverse character of the different portions 
of a hydro-carbon flame, they at least did not corroborate the idea of the re- 
frangibility of its light, varying from its outer to its inner regions. Thus, 
the envelope, the blue conoid, and the bright conoid, which successively enclose 
each other, all produce spectra, having rays corresponding with the lines B and 
G of FRAUNHOFER; lines which differ greatly in refrangibility. It was, moreover, 
a matter of perfect indifference, so far as I could ascertain, from what precise 
point the light was derived. Whether the light were taken from near its outer 
* Philosophical Magazine, vol. xxxii., p. 1038. ¢ Ibid., vol. xxxii., p. 104-105. 
