56 On Anomalous Dispersion in Incandescent Sodium Vapour. 



sion-power of the sodium-flame with increased density of the 

 vapour. While a flame coloured by a sodium-salt, inserted 

 as absorbing medium in the path of the rays, gives two sharp 

 dark lines of absorption in the yellow of the spectrum, when 

 the density of the vapour becomes greater these absorption- 

 lines blend into a single broad band with fainter margins. 

 Now with these bands the dispersion-anomaly appears distinctly 

 visible, while with the narrow absorption-lines, though at any 

 rate present, it is not recognizable, since here it is limited to 

 an extremely small compass in close proximity to the two sides 

 of each absorption-line. 



If we could form a real prism of incandescent sodium- 

 vapour, we should probably be able to observe indications of 

 anomalous dispersion in the narrow absorption-lines, even with 

 less density of the vapour. My endeavours, however, to con- 

 vert the conical flame into a prismatic one by applying plates 

 of glass or mica to its sides, led to no result. Just as little 

 have I hitherto observed dispersion-anomalies in other incan- 

 descent vapours by introducing salts of the metals into the 

 Bunsen burner ; the density of these vapours, and consequently 

 their absorption, is too slight for the method of observation I 

 employed. With improved methods and very dense vapours 

 the same phenomenon as in sodium-vapour will doubtless be 

 obtained. 



To the foregoing I add a remark which, so far as I know, 

 has not yet been enunciated. Those solid and liquid bodies 

 which exhibit for certain groups of rays strong absorption, 

 and for neighbouring groups anomalous dispersion, possess, 

 as I have previously shown, for the same groups a strong re- 

 flecting-power*. 



As it is proved that incandescent gases exhibit anomalous 

 dispersion in the vicinity of the rays which they strongly ab- 

 sorb, it must be assumed, according to the analogy of expe- 

 rience with liquid and solid bodies, that the gases also strongly 

 reflect those rays which they strongly absorb and, conse- 

 quently, emit. A sodium-flame, therefore, would reflect much 

 more strongly rays of the number of vibrations of the D line 

 than any other luminous rays of the spectrum, and therefore 

 show a yellow surface. Experiments in proof of this inference 

 (which presumably would present considerable difficulties) I 

 have not yet instituted. 



Before such experiments are available and the strength of 

 the selective reflections of incandescent gases is in some mea- 

 sure quantitatively determined, it would be precipitate to 



* Conf. Stokes, Pogg, Ann, xci. p. 158 (1854), and xcvi. p. 522 (1835). 



