FLAME SPECTRA AT HIGH TEMPEKATURCS. 
1075 
because the proportion of manganese decreases in the “ pig,” and that contained in 
the slag after the termination of the “ boil ” also decreases, it does not necessarily 
follow that the actual cjuantity of manganese in the slag is diminished. Data are 
wanting which would enable us to decide how much manganese is volatilized, since 
we do not know the absolute quantity of slag and iron. 
Brunner’s analyses do not appear to me to prove that the absolute c^uantity of 
manganese in the slag diminishes or increases during the “ boil,” since manganese 
is not the sole basic constituent of the slag. This question was considered by 
Marshall Watts who, in experiments both at Crewe and at Barrow, always 
observed a difference between the ordinary Bessemer sj^ectrura as seen at Crewe and 
that of spiegel-eisen. The difference, which consisted in a relative intensity of the 
lines, was so great that it was not at first perceived that the spectra were in any way 
the same. At Barrow this difference in the spectrum was not seen, the spectrum of 
spiegel-eisen being identical with that of Bessemer metal, only more intense. The 
ordinary Bessemer spectrum at Barrow was identical with the spiegel spectrum at 
Crewe. The difference between the Barrow and the Crewe spectra was attributed 
by Watts simply to a difference in temperature. It was stated, however, that it 
might have been connected with a difference in composition of the metal operated 
upon. Experiments made on the temperature of the flame showed that at the 
commencement it was below 1300° C., but it gradually rose without reaching 
2000° C. 
From a study of my photographs it appears certain that the whole phenomenon is 
primarily due to rise of temperature, which takes j^lace rapidly and continuously 
during the “boil,” while at the same time an increasing cjuanrity of carbonic oxide 
escapes from the converter. The bath of metal is first heated by the oxidation of 
the manganese and silicon. Such oxidation produces an enormous amount of heat; 
first, because the heat of combustion of these elements is very high ; secondly, 
because the products of combustion are solid, or at high temperatures liquid, and 
carry none of the heat away. This appears to have been first recognised by 
Lieutenant Dutton, U.S.A., in 1871. (‘Chem. News.,’ vol. 23, p. 51.) Then the 
carbon burns and yields a large amount of heat to the metal, the hot metal heats the 
blast which passes through it, and so increases the rapidity of combustion of the 
carbon, which serves again to raise the temperature of the metal. The gaseous 
contents of the converter are carbonic oxide and nitrogen, and within this atmosphere 
the manganese and iron are vaporized, but not oxidized. 
It may be easily understood that the temperature continues to rise until near the 
termination of the “ boil,” because the temperature of the bath of metal increases, 
and consequently the temperature of the blast as it escapes from the metal increases, 
so that the temperature of the combustion of the carbon and also of the carbonic 
oxide is higher. We are, in fact, dealing with combustion under similar conditions to 
those in a Siemens furnace on the regenerative principle. 
6x2 
