258 Dr. W. M. Watts on the Spectra of Carbon. 



There is another element of uncertainty which must not be 

 forgotten. The calculated temperatures and those obtained ex- 

 perimentally by Bunsen are the mean temperatures of the flames, 

 and it is quite possible for one part of a flame to be 1000° C. 

 hotter or 1000° C. colder than the temperature given as the 

 temperature of the flame. The blue cone of a Bunsen gas- 

 flame, from which the carbon-spectrum is obtained, is certainly 

 much colder than the exterior cone of the flame at the same 

 point. 



I have made several attempts to reduce the temperature of the 

 olefiant-gas flame, but have not succeeded in altering the spec- 

 trum at all. Olefiant gas, burnt by means of oxygen in an 

 atmosphere of hydrogen, gives the carbon-spectrum brilliantly 

 with all the fine lines previously described; and a mixture of 

 olefiant gas and steam burns with a colourless flame which ex- 

 hibits the same spectrum. 



A mixture of 2 vols, carbonic anhydride and 1 vol. olefiant 

 gas burns with a barely luminous flame, the blue part of which 

 gives the groups 7, S, e, and / of the carbon- spectrum. The 

 calculated temperature of such a flame is 2016° C. ; but in all 

 probability the temperature is much less, as no allowance is made 

 in the calculation for any refrigerating effect produced by the 

 decomposition of the carbonic anhydride. 



The fusing-point of gold is given by Deville* as 1300° C, 

 and of platinum as 2000° C. The interior blue cone of a Bun- 

 sen-flame about 10 millims. above the jet, which is the part 

 which yields the carbon-spectrum most plainly, is capable of 

 melting gold, but does not melt platinum. It is incapable of 

 fusing steel, which is fused by the outer cone at the same point; 

 and platinum resists the flame at any pointf. We may therefore 

 probably assign to the inner blue cone a temperature of about 

 1500° C. 



The temperature of the flame of olefiant gas and oxygen has 

 not been determined by experiment ; but it can hardly be above 

 2500° C, and we may therefore conclude that the groups 7, 8, e 

 are produced by incandescent carbon between the temperatures 

 of about 1500° C. and 2500° C.J 



In order to determine the inferior limit of the groups £ and 9, 

 a mixture of equal volumes of carbonic anhydride and cyanogen 

 was made; it burnt with a violet flame of small intensity, yield- 

 ing the carbon-spectrum, including the group 6 and the bands 



* Lecons sur la Dissociation, p. 284. 



t A fine platinum wire, which could not be fused in any part of a Bun- 

 scn-ilame, was easily fused at one point in an ordinary bat-wing gas-burner. 



X The groups 7, 8, e are those observe ft by Huggins in the spectrum of 

 Winnecke's comet. 



