Flame Spectra observed at Swedish Bessemer Works. 77 



This instrument was a Hofmann's " spectroscope a vision directe 

 avec six prismes," from Paris, consisting of a collimator, six prisms, 

 a telescope, one eye-piece, and a side tube with a scale. The scale 

 is lighted by a lamp, and the degrees of the scale are reflected by the 

 first prism into the eye- piece, which can be moved by a screw so that 

 the different parts of the spectrum may appear in the field succes- 

 sively. 



To show the excellence of the instrument, the following facts may 

 be noted. It divides the line b of the sun spectrum. The distance 

 between Li a (6705) and Na a (5895) was 41°. The same band 

 showed on the scale the following values : — 42*3, 42*0, 42'5, 42*0, 42*0, 

 42*2. By means of an interpolation curve, the wave-lengths of this band 

 were determined to be as follows : — 5351, 5354, 5350, 5354, 5354, 5352, 

 The interpolation curve was made by numerous observations on the 

 lines of the sun spectrum, the wave-lengths of which were taken 

 from Angstrom's atlas. By means of this instrument all the spectra, 

 except those of Hagfors and Nykroppa, were observed. The spectra 

 of Hagfors and" Nykroppa were examined by a spectroscope of 

 General Wrede's construction, belonging to the Royal Swedish 

 Academy of Science, which instrument proved, however, to be less 

 suitable for my investigations. 



During the Bessemer "blow "there are, as is well known, three 

 periods. 



The first is the so-called "slag period," during which in all cases 

 there is only a bright continuous spectrum visible. Just at the begin- 

 ning of the second period, the " boil," there appears at all the works 

 under my observation, with exception of Langshyttan, first the sodium 

 line, and then a beautiful band- spectrum, which gradually becomes 

 more distinct. It attains its full brilliancy after the entrance of the 

 third period or " fining stage," and then its brightness gradually fades 

 until, upon complete decarbonisation, it vanishes. It is this spectrum 

 that in the following pages I propose to designate the " Bessemer 

 spectrum," and to particularise the Langshytte, Domnarfs, Nykroppa, 

 Hagfors, or Ulfshytte spectrum according to the name of the works 

 where the spectra were studied. 



At Langshyttan, as at several other works, the flame seems during 

 the transition from the first period to the " boil " to assume a more or 

 less intensely blue colour which soon vanishes, after which the colour 

 becomes more of a milky white appearance. 



This blue flame gave at Langshyttan a beautiful spectrum, which 

 appeared suddenly and almost immediately in its full brightness, 

 after which it rapidly lost its colour and changed to a common 

 Bessemer spectrum. This I propose to call the spectrum of the blue 

 flame, and to describe it separately. At Ulfshyttan the metal 

 " blown " contained a comparatively very large percentage of silicon 



