TRANSACTIONS OF SECTION B. 299 



&c, containing the iron protoxide resulting from the oxidation of the iron sulphide, 

 combined as silicate with the siliceous fluxing materials present in the bath ; and 

 underneath this, the regulus or remaining unburnt sulphides, containing in an ap- 

 proximately known state of concentration the more valuable metals derived from 

 the metalliferous substances operated upon. It was, however, necessary for the 

 practical application of the theory that sufficient heat should be developed during the 

 combustion of the iron, zinc, or other less valuable sulphides to keep the materials 

 molten during the operation. The temperatures of combustion of various sulphides, 

 calculated from known data, approached the maximum temperature attainable by 

 the combustion of coal, and this inspired a considerable amount of confidence. In 

 the case of iron pyrites these calculations are only rough approximations, as the 

 latent heat of sulphur vapour is not known. It was found that when thus 

 treating cupreous pyrites, the order in which the elements became oxidised was as 

 follows: — 



1. Zinc and iron. 



2. Sulphur. 



3. Lead and copper. 



The above reactions find a parallel in the elimination of the metalloids from cast 

 iron by Bessemer's process, in which silicon and carbon and then phosphorus and 

 manganese are successively burnt out of the crude metal. Parallel analogies also 

 exist between the process of puddling and English copper smelting; where the 

 oxidation proceeds but slowly, and the necessary heat is obtained by the burning 

 of coal. 



The foregoing conclusions have been verified experimentally ; full particulars 

 thereof will be found in papers brought before the Society of Arts, February 12 

 and April 30 of this year. 



The spectroscopic observations taken by Dr. W. M. Watts during the course of 

 these experiments were valuable and interesting, and I am indebted to him for the 

 following information : — 



In the experiments at Penistone two spectra were observed ; the first, that 

 given by the flame from the charging door of the cupola in which the pyrites was 

 melted ; the second, produced by the blast of air through the molten protosulphide 

 in the converter. The cupola-spectrum was shown by direct comparison with the 

 spectrum of a flame coloured by lead chlorate to be mainly due to oxide of lead, 

 but contained besides some few of the lines which appear to be proper to the 

 converter-spectrum. Analysis showed that the lead present in the ore was almost 

 entirely volatalised during the preliminary melting of the ore, the molten proto- 

 sulphide charged into the converter containing only - 8 per cent. lead. The 

 converter flame gives a brilliant spectrum extending from the lithium line some- 

 what beyond the thallium line, which is usually present. Its most marked feature 

 is the presence of four bright red lines about equally spaced between the sodium 

 and lithium lines. Their wave lengths, as far as at present known, are approxi- 

 mately 5,999, 6,151, 6,320, and 6,466, besides a fainter line at 6,110. These lines 

 are not those of any known spectrum. The way in which the flame is obtained 

 suggests the theory that they are sulphur lines. When sulphur is burned in air or 

 oxygen the spectrum obtained is entirely continuous, and even if air be bubbled 

 through boiling sulphur no lines are obtained. Two spectra of sulphur obtained by 

 the electric discharge through a vacuum tube containing vapour of sulphur have 

 been described, but neither contains these four red lines. The spark with a 

 Leyden jar in a current of sulphur dioxide at the ordinary pressure yields a 

 Bpectrum (at present under investigation) apparently not previously described, in 

 which, however, the red lines are altogether different from those of the converter- 

 spectrum. The constancy with which these four red lines are associated together 

 seems to preclude the possibility of their being due to different substances, other- 

 wise the most refrangible fine might be due to lead. No lines of copper were 

 observed except in the fourth experiment, in which all the lines except those of 

 sodium disappeared about six minutes before the turn-down. When in this ex- 

 periment, towards the end of the blow, the subsulphide of copper began to burn, a 

 splendid emerald green flame suddenly appeared, and all the lines except those of 



