296 Prof. R. W. Wood on the Dispersion of Sodium 
after some of the vapour has condensed there, the spectrum 
straightens out, and the dispersion presently occurs in the 
opposite direction, showing that the base of the prism is 
always against the heated portion of the tube regardless of 
gravity. In a tube 25 cms. long, 10 cms. of which are heated 
red-hot on all sides by means of an iron wire carrying a 
heavy current, a mass of sodium vapour may form of sufficient 
density to give a displacement of several thousand helium 
(Ds) fringes ; notwithstanding this, practically no trace of the 
vapour can be detected beyond the heated portion. This 
condition can be maintained for an hour or more, owing to 
the slowness of the diffusion. If mercury is heated under 
similar conditions, it is impossible to obtain shifts of more 
than a fringe or two. Potassium distils almost in a flash to 
the colder parts of the tube. The other alkali metals have 
not as yet been investigated. The vapour of sodium appears 
to possess a cohesion almost as great as that of aliquid. Hven 
in a vacuum-tube it appears to have a free surface, while in 
an atmosphere of hydrogen it can be dipped up on a wire, a 
vigorous shaking of which is insufficient to dislodge it or 
cause its dissipation. Anumber of very interesting experiments 
with the vapour have already been made, and others are in 
progress at the present time. They will be reported in a 
subsequent paper, as for the present we are only concerned 
with the fact that a homogeneous cylinder of very dense 
vapour terminated with sharply defined ends can be formed 
in an exhausted tube. 
Anomalous dispersion occurs at the absorption-lines which 
belong to the principal series, being very strong at the D 
lines, feeble at the first pair of ultra-violet lines (WX=3303) 
and almost imperceptible at the second pair (A=2852), as is 
shown in fig. 2. If the vapour is of considerable density, 
Fie, 2. 
the D lines run together into a single absorption-band, and 
the dispersion can be represented throughout the complete 
range of wave-lengths as resulting from electrons of a single 
period, for the ultra-violet bands only affect the refractive 
index in their immediate vicinity. These bands make their 
appearance in succession as the density of the vapour in- 
creases, the second not appearing until the band at the D lines 
