262 Dr. W, M. Watts on the Spectra of Carbon. 



of the flame (which is, of course, not really the case), we obtain 

 as the temperature of the flame 1700° C. instead of 2339° C. 



Mr. Bamsbottom has kindly placed at my disposal the result 

 of an experiment made at Crewe to determine the " heat of the 

 Bessemer-flame, in which it was found that on exposing a bar of 

 cast iron (quality not stated), 1J inch in diameter, to the action 

 of the flame at a distance of about 12 inches from the mouth of 

 the vessel, it began to melt in about 5J minutes, the iron drop- 

 ping off in small globules at the rate of about 30 or 40 per 

 minute. A bar of wrought iron exposed in a similar manner for 

 about, six minutes at the end of the blow did not melt." 



We may therefore conclude that the temperature of the Bes- 

 semer-flame lies between 1000° C. and 1500° C. It is worthy 

 of remark (since it throws light on the question whether the 

 carbon-spectra are to be regarded as produced by carbon in the 

 gaseous state or not) that the Bessemer-spectrum contains the 

 lines of iron. There is probably as much difficulty in suppo- 

 sing the existence of iron-vapour below 1500° C. as in supposing 

 the existence of carbon-vapour at the same temperature. 



The Bessemer-spectrum is either due to carbon or to carbonic 

 oxide. If it be produced by carbon, w r e are compelled to admit 

 the existence of two spectra produced by carbon at the same 

 temperature; for the Bessemer-flame lies between 1000° C. and 

 1500° C, and the gas of the vacuum-tube is below 1500° C. 

 If we assume that the Bessemer-spectrum is due to carbonic 

 oxide, we have to explain why in the Bessemer-flame carbonic 

 oxide gives a spectrum consisting of bright lines, and in the 

 carbonic-oxide flame a continuous spectrum. Deville* has shown 

 that the carbonic-oxide flame varies in temperature from about 

 1000° C. at the top of the flame to a temperature considerably 

 above the fusing-point of platinum, or probably 2500° C. at the 

 blue cone 10 millims. from the jet; so that we have then to 

 admit the existence of two spectra of carbonic oxide within 

 the same range of temperature. It may be objected that the 

 determination of temperature is very uncertain, and that if car- 

 bonic oxide were more intensely heated it would give out the 

 Bessemer-spectrum; and indeed at the highest temperature 

 obtainable from carbonic oxide and oxygen burnt together at 

 the oxyhydrogen jet a faint spectrum does become visible from 

 the blue cone, which Deville has shown to possess the highest 

 temperature; but it is identical with the carbon-spectrum 

 No. 1. The probability is that the compound carbonic oxide, 

 like the compound carbonic anhydride, always gives a conti- 

 nuous spectrum — but that at the extremely high temperature 

 obtained in the experiment mentioned above the carbonic oxide 

 * herons sur la Dissociation!, p. 302. 



