OF FLAMES RESULTING FROM BESSEMER PROCESSES. 
489 
the oxyhydrogen flame spectrum of calcium oxide. It corresponds with a strong 
Fraunhofer line (best seen in Plate No. 4). 
The oxyhydrogen spectrum of rubidium is such that no further lines of this element 
were expected, and none were found. In the case of the calcium oxide, the hand in 
the green (the wave-length of its centre is about 5530) might be expected, and a band 
does occur, but this will be discussed later. 
The line X 4171‘5 was identified only after a prolonged study. Of the many 
elements and compounds examined in the oxyhydrogen flame, none gave a strong line 
of this wave-length. Reference to published tables of arc and spark spectra indicated 
that it might be the strongest of the gallium lines. 
Lecocq de Boisbaudran gives the wave-length of the strongest of the lines of 
this element as 4170‘4 on Angstrom’s scale, or 4171’0 on Rowland’s. The second 
line he gives as 4031‘9 (Angstrom), 4032‘6 (Rowland) ; this latter would lie between 
the two closely adjacent lines in the manganese triplet, and it would, therefore, be 
quite impossible to distinguish it when the manganese lines are so strong. No other 
strong lines are recorded, and this indication of the origin of the line could he 
confirmed only by analyses of the materials composing the lining of the converters, 
and of the “basic” earth and metals contained in the charge. By spectrographic 
analysis the same line was found in the oxyhydrogen spectrum of the crude iron 
(“mixer metal”). By chemical analysis the metals forming sesquioxides were freed 
from manganese, and the sesquioxides in this condition showed by spectrographic 
analysis the two strong lines, that with wave-length 4032'7 being weaker than 4171’5. 
'Ihe presence of gallium was jiroved by separating the pure oxide from a weighed 
quantity of metal, and determining the amount present. The purity of the oxide 
separated was established by spectrographic analysis (‘ Proc. Royal Society,’ vol. 60, 
pp. 35 and 393). 
Of the lines next in order, ten are identical with lines in the spectra of flames from 
the “ acid ” process and with lines in the oxyhydrogen flame sj^ectrum of iron and its 
compounds. It must be remarked, however, that we now meet witli notable 
differences in the relative intensities of these lines in the various Bessemer and 
oxyhydrogen flame spectra. 
The differences, we believe, result wholly from differences in temperature, and our 
reasons for this we propose now to state. 
The iron lines in this region of the spectrum ol the oxyhydrogen flame appear on 
the following table, with varying intensities :—■ 
3 R 
VOL. CXCVI.-A. 
