OF FLAMES EESULTING FROM BESSEMER PROCESSES. 
487 
calcium oxide were seen from the very first appearance of flame and foi' aljont two 
minutes of the “ blow” ; but the lithium and potassium lines a})pear rather more than 
a minute after the beginning. On Plate No. 17, which is a very complete series of 
thirteen photographs taken during a “blow” lasting 15 minutes, and commencing 
directly the converter was turned up, it was recorded that the red lines of lithium 
and potassium, with the yellow line of sodium, were plainly visible with the direct 
vision spectroscope as soon as the shower of lime dust cleared away after the “blow” 
commenced, that is to say, from the first appearance of the flame. 
The bands of manganese were seen very early, but they were not sharp during the 
first 4 minutes; this refers to the bands in the green. 
There can be no doubt that the immediate production of a flame is caused not by 
the oxidation of material in the bath, but by carbonaceous matter in the lining of 
the vessel; that its luminosity is due partly to the volatilisation of alkalies, but 
certainly, anterior to this effect, by the ignition of lime dust carried up Ijy the blast 
into the flame. 
Characteristics of the Spectra of the “Basic” Bessemer Flame. 
A flame spectrum is observed shortly after the commencement of the “ blow,” 
consisting of a band of continuous rays of the red and green hands of calcium oxide. 
One minute later the red lines of lithium and potassium are seen with the lines of 
sodium. There are about fifty-three lines of iron as in the flame from the “ acid ” 
process, hut the lines are of lesser intensity, and some of those near to, hut more 
refrangible than, 430G'7 (the solar line G) are absent. 
A number of lines which do not occur in the spectra of the flames from the “acid” 
})rocess have been observed in most, if not all, of the spectra of the “ basic” process. 
They have been traced partly to elements present in the lime, but the lines are chiefly 
caused by elements introduced with the iron, and are hitherto unrecognised con¬ 
stituents of the iron ores from which the crude metal is smelted for use in the 
Bessemer process. Rubidium and csesium undoubtedly accompany the lime, while as 
certainly gallium, copper, and silver, and some rubidium, enter with the iron. Lead 
lines were not observed in the flame, neither were nickel nor cobalt lines. 
Anotlier feature of these spectra is the less prominent ap})earance of the manga¬ 
nese bands, and lines of manganese, with the greater intensity of the lines of 
potassium. 
The description of six spectra on Plate No. 14 (see p. 485) shows the increase in 
temperature of the flame with the progress of the “blow” until that period when 
the flame drops. 
