1904-5.] On Prof. Seeliger's Theory of Temporary Stars. 529 
The bright bands showed two maxima, except Hy, where 
three were noticed. We know, however, from the observations of 
the spectrum of Nova Persei, that the Hy band has a specially 
complex structure, being overlapped by another band on the less 
refrangible side. The third maximum noticed in Nova Aurigae 
may therefore not belong to the Hy radiations at all. Accepting 
this not improbable supposition, we conclude that in Nova Aurigae 
the bright bands showed two maxima with displacement of-0 - 6 
and +6*3 tenth-metres, whereas the absorption-band exhibited a 
maximum in form of a bright line at —9*6 tenth -metres. If we 
assume a common motion of the whole system of about —1*6 t.m., 
the reduced motions of the maxima of the bright band would be 
■h 1*0 and +7*9 t.m., and that of the maximum in the absorption- 
band - 8’0 t.m. Now if we draw the observed intensity-curve 
corresponding to the arbitrary scale of fig. 6, the positions of the 
three maxima will be indicated by the three arrows above the 
curve. We notice, therefore, that the relative distances between 
these maxima are very closely represented by the theoretical curve 
in our diagram. At this stage we may submit the theory to a 
further test, which in my opinion goes far to show its probability. 
Doubtless the presence of the continuous spectrum has a decisive 
influence on the appearance of the bands, whose character on the 
more refrangible side is mainly determined by the absorption-band 
which is caused by gaseous matter moving between the in- 
candescent star and the observer. But we know from observations 
that in Nova Aurigae, as well as in Nova Persei, the continuous 
spectrum has gradually faded away in such a degree that the 
star in its last stages of luminosity was almost entirely reduced to 
its gaseous emissions. Seeliger’s hypothesis explains this course 
of events quite naturally. We have only to consider that an 
incandescent solid or liquid radiates heat more freely than a gas, 
and also that the brilliance of the star nucleus is confined to a 
shallow surface layer whose energy will be rapidly dissipated. 
We may ask : What becomes of the band shown in fig. 6 after the 
continuous spectrum has disappeared 1 ? Considering that the 
former absorptions will now have become radiations, the combined 
spectrum of the bands will be represented by an intensity-curve 
such as is showrn in fig. 7. Here, again, our want of knowledge 
PROC. ROY. SOC. EDIN. — VOL. XXV. 34 
