1904-5.] On Prof. Seeligers Theory of Temporary Stars. 541 
gradually become an emission-line, and, as such, might enhance the 
intensity of the violet edge of the emission-band. We should then 
notice those peculiar finger-post structures (fig. 17) which are 
so prominent features in the later spectrograms of the Lick 
Observatory (see L.O. Bulletin , No. 8). 
There is a good reason for the longer persistence of the less re- 
fracted absorption-line. The more rapidly the atmosphere expands, 
the more quickly will the absorption-band thin out and disappear. 
But since the more refracted band is due to the more rapidly ex- 
panding atmosphere, we may naturally infer that its existence 
A 
4 6 cltvcL 1 ^ 
must be of shorter duration than that of its neighbour, which is 
caused by the less expanding gases. 
The double-star hypothesis, which apparently explains in a 
satisfactory way some of the peculiar spectral features of Nova 
Persei, may also assist us in understanding more fully the peculiar 
variability of the star’s light, specially noticed during the first 
stages of development. In an earlier paper (Astronomische 
Nachrichten , Nos. 3822-3) I have attempted to show that the 
principal features of this variability may be explained by a rotation 
of the star round an axis. I have there emphasised the fact that 
by the more or less one-sided collision the star’s superficial layers 
must be melted unequally, the liquefaction reaching down into 
