1904-5.] On Prof. Seeligers Theory of Temporary Stars. 525 
sides by broad and hazy absorption-lines, these duplex lines being 
projected on the continuous photospheric spectrum. 
This consideration of the conditions prevailing on a star 
whose atmosphere is rapidly expanding leads already to con- 
clusions with regard to the character of its spectrum which are in 
satisfactory agreement with the principal and most important 
feature of the Nova spectrum. Our conclusions require only some 
further modification through the existence of the rotating ring of 
luminous matter we have considered before. Great significance 
must be attached to the fact that the same type of spectrum must 
appear under all circumstances, whatever may be the relative 
positions of star and observer. Hence the remarkable uniformity 
of the spectra of all the Novae appears to be capable of a simple 
explanation. 
According to the views here expressed, the described phenomena 
should occur in a certain sequence which deserves careful attention. 
The immediate effect of the collision being incandescence of the 
star’s surface, the spectrum of the star, at the moment of the 
catastrophe, will be purely continuous. Subsequent to this 
stage, which is probably of short duration, we have the 
development of the expanding atmosphere, which impresses its 
existence on the spectrum only after the expanding gases have 
cooled below the temperature of the surface. At this stage broad 
and diffuse dark lines, strongly displaced towards the violet, make 
their appearance. Some further time will elapse, however, before 
the atmospheric halo round the star has sufficiently expanded 
to render its bright lines visible against the luminous background 
of the continuous spectrum. Now this order of events deduced from 
the theory seems to be confirmed by certain observations. In the 
case of Nova Persei, thanks to its timely discovery by Dr Anderson, 
we w'ere fortunately permitted to watch the celestial catastrophe 
almost from its very commencement. When the spectrum 
was first viewed here in Edinburgh by the Astronomer- Eoyal for- 
Scotland on the early evening of the 22nd February 1901, it 
certainly appeared to be purely continuous. A few hours later,, 
however, I noticed distinctly faint dark bands, one of which agreed 
in position with the absorption-band afterwards noted on the 
violet edge of the bright E-line. On that night emission-bands- 
