Dr. W. M. Watts on the Spectra of Carbon. 259 



of cyanogen. As the temperature calculated for the cyanogen- 

 flame agrees closely with the experimental result, we are probably 

 justified in accepting the calculated temperature in this case also 

 as not very far from the truth, and may therefore conclude that 



6 begins to be visible about 2200° C. Platinum and steel are 

 easily fused in the flame of cyanogen burning in air. 



The temperature of a gas ignited by the electric discharge de- 

 pends upon the resistance and upon the quantity of electricity 

 transmitted in each spark ; and this may be increased either by 

 increasing the condensing surface, or by increasing the tension of 

 the electricity at discharge. This tension depends upon the 

 nature of the gas : thus the spark passes with great ease through 

 hydrogen, with more resistance through carbonic oxide, carbonic 

 anhydride or oxygen, and with extreme difficulty through cya- 

 nogen. But for one and the same gas the tension at discharge 

 and resistance experienced are increased by increasing the den- 

 sity; and the heat produced thus increasing more rapidly than 

 the quantity of matter to be heated, the temperature rises*. 

 Hence we understand why the groups f and 6 are added to the 

 spectrum of the spark in carbonic oxide when the density of the 

 gas is increased, and why the spark in the gas cyanogen, which 

 offers so great resistance, always gives a spectrum containing f and 

 6. We have also the explanation of the fact that a vacuum-tube 

 containing either the dense vapour of naphthalin, or the badly 

 conducting gas cyanogen, gives -always the spectrum of carbon 

 belonging to the high temperature, although the pressure of the 

 gas is only a few millimetres. 



It is impossible to assign any temperature as the superior 

 limit of this first form of the carbon-spectrum which shall have 

 any meaning, or to guess with any probability at the temperature 

 of the condensed spark. It cannot be less than 10,000° C. ; but 

 the temperature calculated for the flame of cyanogen in oxygen 

 (without doubt the hottest flame known) can hardly be trusted. 

 The carbon-spectrum No. 4 may then roughly be said to be due 

 to incandescent carbon above 10,000° C. 



It has been shown that carbon at 1500° C. gives the first form 

 of carbon-spectrum, and that the same spectrum is given by the 

 electric spark in either carbonic oxide or defiant gas at the ordi- 



* I have repeatedly observed this increase of resistance in the experi- 

 ment on condensed gases. The spark which passed with ease in carbonic 

 anhydride at the ordinary pressure could hardly be got through the gas at 



7 atmospheres' pressure, while there was no perceptible increase in the re- 

 sistance afforded by hydrogen when the pressure was increased to 9 atmo- 

 spheres; and the spark which passed with ease through 7 millims. in 

 hydrogen at 9 atmospheres' pressure would hardly pass through 5 millims. 

 in cyanogen at the ordinary pressure, and through only a fraction of a mil- 

 limetre in cyanogen at 4 atmospheres' pressure. 



