the Radiation of Heated Gases. 475 



Origin of the Continuous Spectrum. 



There remains to be considered the continuous light ssen 

 when the sodium is first volatilized, and before the broad 

 D emission-line appears (fig. 3, a). After a number of ex- 

 periments, made under various conditions as to the parity of 

 the coal-gas or nitrogen employed and density of the sodium 

 vapour, I conclude that this radiation is really due to free 

 sodium, and that it always appears when the vapour is above 

 a certain density. It is well known that dense sodium vapour 

 gives a remarkable banded absorption-spectrum *. In the 

 experiment alluded to, I have studied the continuous emission- 

 spectrum in relation to this banded absorption, and find that 

 they are intimately connected : thus, when the sodium is first 

 volatilized and a bright glow fills the tube, the vapour appears a 

 splendid violet colour by transmitted light ; it is in fact opaque 

 to all rays except the violet ; gradually, however, green rays 

 begin to be transmitted (fig. 4, a) , then red (fig. 4, b) ; the dark 

 space between the green and the violet is now seen to be made 

 up of a large number of black lines (close together with the low 

 dispersion employed), these rapidly decrease in intensity as 

 the vapour diffuses along the tube and becomes less dense, 

 while the black band blotting out the yellow assumes the 

 usual appearance of the wide D absorption-band f . The 

 continuous emission glow persists through these progressive 

 changes in the absorption, but becomes gradually fainter, 

 and as the last traces of the absorption-lines in the blue dis- 

 appear it fades away almost entirely, leaving behind, so to 

 speak, the broad bright emission-band at D, the exact reverse of 

 the absorption-band, except for the black line filling up the 

 centre, due to relatively cool vapour in the nearer part of the tube. 



It seems, therefore, that dense sodium vapour, like iodine 

 and other coloured vapours giving banded absorption-spectra, 

 emits when in this state light of all wave-lengths, and that 

 the change to discontinuous emission is determined by a 

 reduction of density. There are indications, however, that 

 at higher temperatures the continuous light would give place 

 to bands corresponding with the absorption-bands, just as 

 iodine is said to give a banded emission spectrum when heated 



* Vide Roscoe & Schuster, Proc. Roy Soc. xxii. p. 362. 



t A curious phenomenon may be seen when a current or gust of gas 

 is allowed to drive the dense vapour along the tube. The colour of the 

 vapour seen by transmitted light suddenly changes from green to a 

 splendid ruby-red. This is not due to a change in the character of the 

 absorption-spectrum, but _aay be explained as an effect of refraction : 

 the sodium vapour, being blown out more in the centre than along the 

 sides of the tube, acts like a prism, refracting the blue and green rays 

 radially from the centre, the red alone — b3ing least deviated — passing 

 out at the end of the tube. This effect is more obvious with hydrogea 

 than with nitrogen. 



