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



used for experiments on gases under pressure, shows at the or- 

 dinary pressure only a faint continuous spectrum ; at two atmo- 

 spheres' pressure the spectrum is much brighter but still conti- 

 nuous ; and at pressures between seven and ten atmospheres' the 

 spark passes with difficulty, and the spectrum shows a number 

 of bright bands which agree in position with the lines 76, 99, 

 103, and 106 of fig. 4Z>. They differ in character, however, 

 being bands instead of fine lines, thus bearing the same relation 

 to the fine lines obtained from carbonic anhydride at the ordinary 

 pressure as the expanded lines of hydrogen do to the fine lines 

 obtained from a hydrogen vacuum-tube. These bands are ob- 

 tained also in the spectrum of the condensed spark in the vapour 

 of amylic alcohol. 



The spectrum of the direct discharge in a tube containing 

 hydrogen of a few millimetres tension only and a trace of methyl- 

 oxalic ether is faint, but contains the lines k, I, m of the second 

 carbon-spectrum ; but when by warming the tube the ether is 

 volatilized, the spark passes only in brilliant flashes, and the 

 spectrum then contains lines 34, 75, 85-90, 99, 103, 106, 120, 

 125, and 140 of the fourth carbon-spectrum again as bands. 



This fourth spectrum, obtained from carbonic oxide and car- 

 bonic anhydride, may either be due to carbon, or to carbonic 

 oxide, or to carbonic anhydride. It is, of course, not the spec- 

 trum of oxygen. I believe it to be due to carbon-, but I have 

 not been able to obtain such complete evidence as is afforded for 

 the spectra Nos. 1 and 2 in their production from different car- 

 bon-compounds. Thus I have not been able to obtain this fourth 

 spectrum from a compound of carbon with hydrogen alone ; the 

 condensed spark in cyanogen at the ordinary pressure gives, 

 however, together with the carbon-spectrum No. 1 and the ni- 

 trogen-spectrum of the second order, the- lines 34, 56, 76, and 

 103 of the carbon-spectrum No. 4. This conclusion (that the 

 spectrum is really due to carbon itself) seems to be supported by 

 the fact that, when this spectrum is obtained from either carbonic 

 oxide or carbonic anhydride, there is always a perceptible deposit 

 of carbon ; since if it were due to carbonic oxide we should not ex- 

 pect to have carbon deposited in either case ; and if it were due to 

 carbonic anhydride, though carbon would be set free from the car- 

 bonic oxide, there would be none from carbonic anhydride itself. 

 It would appear that carbonic oxide is more easily decomposed 

 than carbonic anhydride, either into carbon and carbonic anhy- 

 dride, or into carbon and oxygen ; so that at the low temperature 

 of the direct discharge carbonic oxide is decomposed and gives 

 the carbon- spectrum No. 1, while carbonic anhydride resists de- 

 composition. If the temperature of the spark be increased either 

 by the intercalation of a Leyden jar or by increasing the density 



