MECHANICS AND USEFUL ARTS. 15 



care now bestowed upon keeping the percentage of carbon in tlie 

 spiegeleisea to a uniform or at least to a correctly ascertained 

 amount, and regulating the quantit}^ of spiogel employed by ex- 

 act calculation. The accidental irregularities and differences of 

 hardness between the different charges have thereby been lessened 

 to a very considerable extent, and an increased reliability has by 

 these means been given to the Bessemer process. 



The spectrum pointed out by Prof. Lieliegg ("Austrian Gazette 

 for Mining and Metallurgy") belongs to the flame of carljonie 

 oxide. It can be seen in the flame escaping from the mouth of 

 the converter during the preliminary operation of heating this 

 vessel with coke only; in that case the lines referred to are very 

 faint, and it requires some practice, or knowledge of the precises 

 spots in the spectrum where these bright lines should be looked 

 for, to discover them. During the first period of the Bessemer 

 process the spectrum is very faint; the yellow portion is almost 

 invisible, and even the sodium line is missing; the blue and ])ur- 

 ple portions are extremely faint. The absence of the sodium line 

 can be accounted for only by the consideration that there is no 

 real flame formed by incandescent gases escaping from the con- 

 verter at that early stage, but only a mass of sparks carried by 

 the nitrogen from the blast, the oxygen of which remains in the 

 converter, combining with silicium. As the flame gradually ap- 

 pears in the centre of the volley of sparks, the spectrum widens 

 and shows yellow light, until suddenly the sodium line in the j'el- 

 low field becomes visible, at first, only for moments, as a flashing 

 bright streak, and after less than one minute as a constant and 

 clearly defined line. The appearance of the sodium line marks 

 the commencement of the decarburization, although this line does 

 not belong to the charge of iron at all, but rather to the accidental 

 presence of sodium compounds in very minute quantities. It is, 

 therefore, only indirectly connected with the combustion of car- 

 bon ; that is, the appearance of the sodium line is a signal of the 

 completion of the continuous spectrum, and this continuous spec- 

 trum belongs to the combustion of carbon. As soon as the sodium 

 line has taken a steady and permanent appearance, the character- 

 istic lines of the carbonic oxide may be looked for in the greenish- 

 yellow, in the green, and in the purple field. In each of these 

 three fields one bright line becomes clearly visible at that time. 

 As the flame increases in size and brilliancy, the spectrum comes 

 out more and more clearly; bright lines increase in number in 

 each of the first-named three fields, and ultimately, at the height 

 of the process, some bright lines show themselves in the red, and 

 occasionally, also, in the blue field ; the green field of the spec- 

 trum, however, is the real point of observation in practice, as in 

 this the lines are most clearly visible, and in it they appear first 

 and disappear last. The s])ectrum, as a whole, is by no means 

 steady or constant, but its fluctuations do not displace any of the 

 bright lines ; they only alter the i)ackground or the continuous 

 spectrum ujDon which the}' appear. After the "boil "the maxi- 

 mum intensity is reached ; and at that stage, and only with very 

 hot charges, a bundle of bright lines appears in the bluish-purple 



