OF FLAMES RESLLTINCI FROM BESSEMER PROCESSES. 
495 
The Spectrum of the Over-blow.” 
During the earlier part of the “ over-l^low ” the flame is very short and not very 
luminous ; the light emitted acts but feebly on a photographic plate. There is more 
fume expelled during this stage than during the first or carbon period of the “ blow,” 
and the quantity increases as the “ over-blow ” proceeds ; the flame during this time 
increases in size and lirilliancy, and towards the end emits a brilliant golden-yellow 
coloured light. There is a notable increase in the quantity of fume expelled after the 
‘‘over-blow” has proceeded for 2 minutes. 
The spectrum of the flame during these 2 minutes is very feeble, but during the 
latter 3^ minutes it is much stronger. It extends from the red end of the .spectrum 
to about wave-length 4000, in the violet; it is strongest in the green about wave- 
length 5500, but fades rapidly towards the violet. 
Tlie following lines have l)een observed in the photographs :— 
APave-length. 
5893. A line present in all the .spectra, due to sodium. 
5432. Sharp, observed only on one plate. No. 16. 
5394. Sharp, obseiwed on two plates. Nos. 16 and 17. 
^ 0 ^ 4 ' present in all spectra due to potassium. 
4034. "j 
4033, VLines present in all .spectra, due to mangane.se. 
4030. J 
These lines are not as sharp on the photographs as the corresponding lines in the 
spectra photographed during the finst period of the “ blow.” 
The two lines, Avave-lengths 5432 and 5394, coincide with two lines near the edges 
of the bands in the flame spectra of the manganese and its compounds. It was not 
apparent at first why these lines should be present in the spectra of some flames and 
not in the spectra of others, and they were not observed in the eye observations made 
during our first visit to Middlesbrough. In August, 1898, on a second visit to the 
same works, eye observations were made with the object of solving the problem. 
Three vessels. Nos. 2 , 3, and 4, were in use on the day in question, and it v/as found 
that these two lines were strongest in the spectrum of the flame issuing during the 
“blow” from vessel No. 4, and also that the edge of the strongest green band was 
also seen. 
The lines were sharp, and, when using a very narrow slit, stood out clearly on the 
strong continuous spectrum. The two lines were stronger in the spectrum of the 
flame issuing from No. 3 vessel than in that from No. 2 vessel. They were observed 
in two or three “ blows ” in each vessel and with similar results each time. 
