(Old Ahsorption Spectra of Sodium Vapour. 373 



eelges ot' which are broken up into smaller bands appearin<;" 

 between 507i) and 5200, and a series of twelve bands, each 

 of which is divided into smaller bands comino- into view 

 between 4780 and 4600. The wave-lengths of the principal 

 lines in the fluted si)ectrum have been carefully measured bv 

 comparing them with standard lines in the iron spectrum. 



A fluted absorption- spectrum makes its appearance in the 

 red and orange portion, when the density ot' the vapour is 

 such as to give the o-reen-blue flutino-s at their best. This 

 spectrum has not been photographed and measured at the 

 present time owing to the difficulty of getting plates suffi- 

 ciently sensitive to the red. The Erythro plates, which were 

 used by one o£ us in securing photographs of this region with a 

 14-foot grating some years ago, are no longer on the market, 

 and the few experiments which we have made in sensitizing 

 our own plates have been only partially successful. This 

 region will be studied as soon as suitable plates can be 

 obtained. 



As the density of the vapour is further increased, the 

 greenish-blue region disappears entirely, and the fine lines 

 can be traced nearly up to w^ave-length 54 from the blue end 

 of the spectrum, and from the red end they are seen to fill 

 up the spectrum^ quite up to the broad D-line absorption, 

 which is now 40 Angstrom units in wddth. On crossing this 

 broad black band the fine lines make their appearance again, 

 which makes it seem probable that the red fluted spectrum 

 crosses the region occupied by the D-lines. A further 

 increase in the density results in blotting out the red, orange, 

 and yellow completely, leaving only a rather narrow green 

 region, the centre of which is at wave-length 5530 and the 

 violet below 4600. The colour of the transmitted light is 

 now a very deep violet. The fine lines can be pushed into the 

 green band from its opposite edges, and traces of them haA^e 

 been found all through it. Finally a very black broad band 

 appears at the centre of the green strip, which in a spectro- 

 scope of small dispersion appears as a rather narrow line. 

 The centre of this band is at wave-length 5530 approxi- 

 mately. We have not been able to photograph it with the 

 grating, owing to the feebleness of the light, and the difficult}' 

 of keeping the vapour at this great density for a sufficient 

 length of time, but its position was determined visually with 

 a grating of about 8 feet radius. This band has been seen 

 by previous observers (Liveing and Dewar), and is described 

 as persisting until the last traces of the fluted spectrum dis- 

 appeared, as the vapour cooled ofl". We cannot understand 

 this, for w^e have only found it when the vapour had the 

 maximum density attainal)le, namely, when a blast-lam}) 



