294 Prof. R. W. Wood on the Dispersion of Sodium 
length been developed which have made possible accurate 
determinations of the absolute refractive index of the metallic 
vapour at various densities, for all wave-lengths comprised 
between the extreme red and the remote ultra-violet. For 
wave-lengths immediately adjoining the D lines the refractive 
index has been found to have a value as high as 1°38, as 
great as that of some liquids, while the dispersion is so 
great even at the position of the D line of helium, that, could 
we form a prism of the vapour giving the same deviation as 
a 60° glass prism, we could by its aid separate a close Zeeman 
doublet by an amount as great as the distance between the 
red and blue of the spectrum formed by the glass prism. The 
results obtained have been applied to the formula, and most 
excellent agreement found. 
The anomalous dispersion of the vapour was first observed 
by Kundt, and has since been studied by Becquerel, Julius, 
Ebert, and others. These investigators have for the most 
part limited their observations to the dispersion produced by 
flames of prismatic form, containing the vapour of sodium. 
Under these conditions the anomalous dispersion is only 
apparent in the immediate vicinity of the D lines, extending 
toa distance up and down the spectrum not much greater than 
the distance which separates the lines. Some four years ago 
it was shown by the writer * that by employing the vapour of 
the metal in glass tubes, it was possible to obtain far greater 
deviations, and to extend the observations on the dispersion 
from the extreme red to the violet. These results were made 
possible by the very remarkable physical properties of the 
vapour, which will be discussed more in detail presently. 
The prisms formed in heated glass tubes are in reality non- 
homogeneous cylinders of the vapour, the density being 
greatest along the heated floor of the tube and least along the 
top. The definition given by non-homogeneous masses of 
vapour is surprisingly good, and relative values of the refrac- 
tive indices can be determined without difficulty; but as no 
means could be found of determining the angle of the 
equivalent prism, it was impossible to do more than guess at 
the absolute values. An attempt made several years ago to 
obtain data by interference methods gave no results, owing 
to the fact that the fringes were immediately obliterated by 
the convection-currents of heated hydrogen. By adopting 
the expedient of heating the metal in a high vacuum, it has 
been found possible to obtain deviations as high as 1500 fringe- 
widths with the D; light of helium, without in any way 
* Wood, Proc. Roy. Soc. (1901) & Phil. Mag. [6] vol. iii., “The 
Anomalous Dispersion of Sodium Vapour.” 
