522 Proceedings of Royal Society of Edinburgh. [sbss. 
Such velocities are of the order of magnitude which would corre- 
spond to the displacements in the spectra of new stars if they 
were to be explained on the basis of Doppler’s principle. Hence, 
if the expansion of the Nova atmosphere is associated with 
enormous surface temperature, the velocities involved must have 
a most profound bearing on the structure of the spectrum. This 
conclusion necessitates an examination of the spectral character of 
lines emitted by gases which form a rapidly expanding atmosphere 
round the incandescent nucleus of the star. It may he well to 
remember at this stage that we are by no means unfamiliar with 
the phenomenon of rapid gaseous expansion at the surfaces of 
celestial bodies. Notable instances are afforded in the solar 
eruptions, where the motions sometimes recorded fall little short 
of the sun’s critical velocity. My opinion on these phenomena, 
more fully expressed in the paper referred to, is that they are 
the inevitable consequences of local changes of temperature in 
the interior layers of the sun. If, for some reason or other, upon 
which I will not enter at present, the temperature of the photo- 
sphere at a certain locality should he raised to that, say, of a 
Sirian star, the conditions of equilibrium over this particular 
area would require that the hydrogen atmosphere of the sun 
should expand to the dimensions of the hydrogen atmosphere of 
Sirius. What we perceive in a solar eruption is therefore, accord- 
ing to this view, the violent transition from a state of atmospheric 
equilibrium at solar temperature to that corresponding to the 
higher Sirian temperature. If the sun were suddenly bombarded 
by a shower of meteors, raising the temperature of the photosphere, 
an inevitable consequence would he the rapid development of a 
protuberance over the place of impact, simply because the atmo- 
sphere would tend to assume that form of thermal and mechanical 
equilibrium which corresponds to the higher temperature of the 
layers underneath. I conclude that the expansion in solar erup- 
tions and that of the atmospheres on new stars are analogous 
phenomena, in both cases due to the tendency on the part of 
the gaseous molecules to assume that state of equilibrium which 
corresponds to the temperature at the surface. If this analogy he 
accepted, and if we take note of the high velocities so often re- 
vealed in the solar gases, we see probably no further difficulty in 
