1871J On Double Spectra. 13 



Leyden jar is included in the circuit. The spark without the 

 Leyden jar is not hot enough to decompose carbonic acid, 

 and gives in that gas only a continuous spectrum. In car- 

 bonic oxide (which seems to be more easily decomposed) it 

 yields the first carbon spectrum, but when the Leyden jar is 

 employed, both gases are decomposed with deposition of 

 carbon, and give the same spectrum. This fourth carbon 

 spectrum (represented in the chromolithograph, Fig. 4) is one 

 of Pliicker's " spectra of the second order," consisting of 

 sharply defined lines, often in pairs, and is no doubt produced 

 by carbon at an extremely high temperature. 



The attempt to ascertain the circumstances under which 

 these four carbon spectra are produced, has led to a result 

 which, if it can be maintained, is of the greatest interest. 



We have seen that the explanation given of the different 

 spectra of hydrogen is that they are produced by the gas 

 heated to different temperatures, and there is no difficulty in 

 conceiving that the particles of a gas may vibrate differently 

 when the gas is differently heated. We have only very 

 rough means of measuring high temperatures, and the de- 

 termination of the temperature to which the gas in a flame 

 is heated can only be approximate. It is true that we 

 can calculate the temperature of a flame from the known 

 amount of heat given out by the gas in burning, but 

 these results are always too high, in consequence of 

 conditions which it is impossible to take into account 

 in the calculation. We have, however, one or two experi- 

 mental determinations of value. Deville and Debray have 

 determined the temperature of the oxyhydrogen flame 

 by heating a mass of platinum to the highest possible tem- 

 perature in a crucible of lime, and then plunging it into 

 water, and determining the temperature to which the water 

 is heated, and find in this way that the temperature is not 

 higher than 2500 C. Bunsen, by a quite different method has 

 found for the same number 2800 C, which, considering the 

 difficulty of the determination, agrees satisfactorily with 

 Deville's result. Bunsen has also found that the tempera- 

 ture of the flame of hydrogen in air is 2024 , of carbonic 

 oxide in air, 1997°, and in oxygen, 3033 , and of cyanogen in 

 air, 3297° C. Deville and Debray have also determined by 

 their method the fusing point of platinum, which they find 

 lies between 1800 and 2000 C. 



The first carbon spectrum is obtainable from flames which 

 cover a considerable range of temperature. It is given by 

 the blue cone of a Bunsen flame, the temperature of which 

 is not high enough to melt the finest platinum wire, and must 



