OF fla:\ifs resulting from. bessei\ier processes. 
505 
may be foimd in the observation of Gouy, that a large excess of air causes the 
spectra of lithium, calcium, strontium, and barium salts to disa})pear, and the 
maximum intensity of their flame spectra is attained before the fiame ceases to exert 
a reducing action. The spectrum of sodium is not affected in the same way, but, on 
the contrary, the intensity of its emitted rays increases rapidly as the flame 
diminishes in reducing power, and is at its maximnm at the instant it ceases to l)e a 
reducing flame. The snbsecpient reduction in brightness is much slower than with 
the other metals. We Ijelieve, therefore, that Gouy’s results, in part, offer a 
satisfactory explanation of the absence of the lines and bands of calcium, and account 
for the feeble character of the other lines and for the changes in the spectrum already 
described (see pp. 485 and 498). 
Fifthly, the spectra of flames from the “basic” process differ from those of the 
“ acid ” jirocess in this respect that tlie manganese bands are almost entirely alisent, 
but lines of elements not usually associated with Bessemer metal, and which have not 
ap])eared in previous photographs, are seen in these. We have proved that the 
difference in the spectra is due in the first place to a difference in the composition 
of the crude blast-furnace iron, which is best adapted to tlie working of the basic 
process. A second reason is the use of charges of “ basic ” materials placed in tlie 
vessel with the metal. The potassium, sodium, lithium, rubidinm, and cmsinm liave 
been traced to the lime ; tlie manganese, copper, silver, and gallium to the ore. 
Other constituents, such as vanadium and titanium, are not in evidence, because 
they do not yield flame spectra, and, furthermore, because they pass in an oxidised 
state along with the phosphorus into the “ basic” cinder. 
(3.) Differences in the Inteusitij of Metallic Lines. 
Iron lines to the number of fifty-three have been measured, as in the spectra of 
flames from the “ acid ” process, but they are of lower intensity, and some of those 
near the position of the solar line G are absent. 
There is a less prominent appearance of the manganese lines and liands, tlie cause 
of which has already been stated ; but tliere is a greater intensity of the lines of 
potassium and calcium, and to these the elements—rubidium, Cfesium, gallium, 
copper, and silver, must be added. 
The variations in intensity of tlie iron lines in the “ basic ” Bessemer flame are due 
to differences in temperature solely. 
There are two lines which in some spectra appear to have become intensified, while 
the calcium band wave-length about 5540 has become greatly enfeelded. It seemed 
as if they had arisen out of the calcium band, but they are not calcium lines. 
They have been observed in many different spectra of Bessemer flames and of iron 
and its compounds in the oxyhydrogen flame, and they are strongest where the iron 
lines are strongest. As no corresjoonding lines have been found in the spectrum of 
any other element, we believe we are justified in attributing them to iron. 
Another line, 5792, was identified with iron in a similar manner. 
3 T 
VOL. CXCVI. — A. 
