1042 
PROFESSOR W. N. HARTLEY OH 
Manchester Proc./ vol. 3, p. 57) stated that he had been engaged during the 
previous year in an interesting examination of the spectrum produced by this flame, 
and had observed a complicated but characteristic series of bright lines and dark 
aljsorption bands. The v/ell-kno\vn sodium, lithium, and potassium lines were most 
conspicuous among many other lines of undetermined origin. In a lecture delivered 
at the Royal Institution (May 6th, 1864) he described the Bessemer flame spectrum 
more fully, and pointed out the existence of lines which he believed to be due to the 
elements carbon, iron, sodium, lithium, potassium, hydrogen, and nitrogen. These 
observations led to the discovery that the exact point of decarburization of the metal 
could be determined by means of the spectroscope with much greater exactitude than 
from the mere appearance of the flame, and for determining the point at which it was 
necessary to stop the blast this instrument was in constant use at Sheffield in 1863, 
and was introduced into the steel works of the London and North-Western Railway 
Company at Crewe (‘ Phil. Mag.,’ vol. 34, p. 437, 1867). 
F. Kohx, in a lecture (‘ Dingler’s Polytech. Journal,’ vol. 175, p. 296) delivered in 
1864, on recent improvements in the Bessemer process, stated that endeavours to 
make spectrum analysis applicable to the Swedish process had not led to any useful 
result. Tujsiner, in 1865 (‘ Dingler’s Polytech. Journal,’ vol. 178, p. 465), stated 
that up to the close of the jirevious year the observations of the flame and sparks 
issuing from the converter and test pieces of the metal, gave better indications than 
spectrum analysis. In 1867, Lielegg made observations on the spectrum of the 
Bessemer flame at the works of the Imperial Southern Railway at Gratz (• Sitzungs- 
berichte der Kaiserl. Akademie der Wissenschaften,’ Vienna, vol. 56, Part II., June), 
which led to the practical application in Austria of the use of the spectroscope to the 
control of the Bessemer process. The spectrum, as described by Lielegg, consists of 
the vapours of sodium, potassium, and lithium, with that of the flame of carbonic 
oxide. Accompanying the latter gas is nitrogen, but no spectrum of this gas appears, 
nor could any spectrum of it be obtained by burning compounds which did not 
contain nitrogen, along with atmosjiheric air. It wms also shown that though carbon 
spectra are obtained by burning hydrocarbons and cyanogen, with air or oxygen, yet 
carbonic oxide yields no such spectrum. Carbonic oxide was found to yield only a 
continuous spectrum when burnt with air or oxygen, or even with nitrous oxide. No 
dark or bright lines were visible. It was represented that the bright lines in the 
Bessemer flame must result from the much higher temperature -which is produced 
by combustion of the heated gas at the mouth of the converter, than wdien laboratoiy 
experiments are performed with the same gas. 
T’he S{)ectruni as it occurs in the “ boil,” and up to the end of the “ fining ” period, 
was described as follows. The figures are arbitrary scale measurements :— 
