Vapour in the Visible and Ultra-violet Regions. 297 
has become three or four times as broad as the distance 
between the lines. The complicated channelled absorption 
spectrum appears to exert no influence on the refractivity of 
the vapour. I feel certain, however, that the channelled 
absorption spectrum affects the refractivity of the medium in 
a small degree, and intend to test the matter in the near future 
by a modification of the method of strize. The channelled 
bands affect the magnetic rotatory dispersion in a powerful 
degree, and this subject is under investigation at the present 
time. The magnetic dispersion is many times greater than 
the ordinary dispersion, as has been found by examining the 
vacuum dispersion-tubes in a powerful magnetic field with 
polarized light, and the whole subject will be fully taken up 
in a subsequent paper. 
The absolute values of the refractive index for different 
densities have been determined with the interferometer for 
monochromatic light of different wave-lengths. Relative 
values (dispersion) have been obtained in the same way, as 
well as by the method of crossed prisms. 
Determination of the Dispersion with the Interferometer. 
The absolute determinations being the more difficult, it was 
decided to begin with an investigation of the dispersion, in 
order to getas much experience as possible with the behaviour 
of the vapour under various conditions. This was a wise 
decision, as was subsequently found, for a number of very 
elusive sources of error were discovered, which were eliminated 
in the final work. 
The general plan of the apparatus is shown in Plate X. The 
interferometer was the form of the Michelson instrument 
designed by Brace. 
The base of the instrument consisted of a heavy tube of brass 
12 cms. in diameter and a little over a metre long. The 
arrangement of the mirrors is shown in the diagram, all but 
one being of speculum metal. Light from an arc-lamp was 
focussed on the slit of the monochromatic illuminator. The 
beam issuing from the slit of this instrument comprised a 
range of wave-lengths about equal to the range between the 
D lines. Near the slit was mounted a helium tube, which 
furnished a light of constant wave-length, and the observations 
consisted in counting the fringe-shifts for the two kinds of 
light when a given amount of sodium vapour was introduced 
into the path of one of the interfering beams. A small 
mirror was mounted in front of the interferometer, which 
deflected a portion of the emergent light into a telescope, the 
remainder of the beam entering a second telescope. By a 
