FLAME SPECTRA AT HCGH TEMPERATURES. 
1073 
The Constitution of the Bessemer Spectrum. 
It will be readily understood from the previous investigations of the flame-spectra 
of iron, manganese, spiegel-eisen, ferro-manganese, silico-spiegel, tool-steel, pure man¬ 
ganic oxide, carbon, carbonic oxide, and cyanogen, that the Bessemer-flame spectrum 
is not characterised especially by the bands of carbon, as would be the case according 
to the views of Roscoe, nor of carbonic oxide according to Lielegg and Kupel™ 
wiESER, nor does it belong entirely to manganese as indicated by the observations of 
Brunner, von Lichtenfels, and Wedding ; furthermore, it cannot be attributed 
chiefly to manganic oxide, as stated by Watts. It is, in fact, a complex spectrum, 
in which the bands of metallic manganese, carbon, carbonic oxide, and cyanogen, 
possibly also of manganic oxide, are superposed ; and the lines of iron and manganese 
occur with those of other elements, such as hydrogen, lithium, potassium, and sodium. 
The hydrogen line (C in the solar spectrum) was photographed only once, and then 
during a snowstorm, when it appeared completely reversed. No absorption bands 
were at any time visible when observations were made upon the flame only. No 
nitrogen bands were seen. No bands belonging to calcium and magnesium oxides 
were visible, nor lines of these elements. No cobalt, nickel, copper, nor chromium 
were detected. The lines beyond the solar line K, which had hitherto not been 
examined, are nearly all lines of iron, as mapped by Cornu and Rowland in the 
solar spectrum, and observed in the arc-spectrum of iron photographed by Kayser 
and Runge. 
Cause of the Non-appearance of Lines at the Commencement and Termination of 
the “ Bloiv.” 
SiLLiMAN detected thirty-three lines in the Bessemer-spectrum; some of Ltelegg’s 
lines were not observed, and others which he did not record were found. Dark bands 
were observed, crossed by bright lines; and it is suggested that the brilliant lines 
tend to make a weak continuous spectrum appear discontinuous, the dark bands being 
merely intervals between the bright ones. The iron spectrum had not been satisfac¬ 
torily identified. According to Silliman's statement, “ the Bessemer-spectrum 
contains yet many mysteries to be solved, among which is the cause of the non- 
appearance of the lines of the spectrum at the beginning and termination of the 
‘ blow.’ ” 
Wedding accounted for the absence of the spectrum at the beginning and termina¬ 
tion of the “blow” by the absolute quantity of the substance volatilized being at 
these times too small to produce a spectrum. (“ Das Spectrum der Bessemer flamme.” 
‘ Zeitschrift fiir das Berg-Hiltten-und Salinen-wesen,’ vol. 27, p. 117, 1869.) He 
based his view upon the fact, recognized by Simmler, that a much larger quantity of 
manganese is required to obtain a recognisable reaction in the flame than that which 
can be detected by the well-known blowpipe test with carbonate of soda. In the 
MDCCCXCIV.—A. 6 X 
