520 Proceedings of Royal Society of Edinburgh. [sess. 
which otherwise might not he capable of bringing the cosmic 
invader under its permanent gravitational sway. 
Still another consequence of importance, however, must be 
drawn from those fundamental considerations by which Professor 
Seeliger was guided in the framing of his theory. We admit that 
the development of heat at the surface of the star must be 
enormous, even granting the most unfavourable assumptions as to 
the tenuity of the impinging cloud. We may safely assume that 
the amount of heat developed during the bombardment may have 
exceeded many times that expended by the sun during a corre- 
sponding time. This fact seems to warrant the conclusion, not 
only that the surface of the star is rapidly liquefied, but also that 
from this surface of molten lava an incessant escape takes place 
of molecules with extremely high velocities, leading to the 
formation of an expanding incandescent atmosphere of vapours 
and gases. This unquestionable fact of an expanding atmosphere 
has, so far, not been considered in theories of temporary stars. 
Is there reason for neglecting the influence of its motions on the 
appearance of the lines in the Nova spectrum 1 There can be little 
doubt that the gaseous molecules escaping from the liquid surface 
of the star would tend towards a state of equilibrium such as is 
presented in the gaseous envelopes surrounding the photospheres 
of ordinary stars. The height of this ‘ atmosphere ’ is determined 
by gravitation on the one hand, and by the surface temperature 
on the other. If, for instance, we assume the mass and radius of 
the star to be equal to that of the sun, and its surface temperature 
to that of the solar photosphere, then the atmosphere would most 
probably assume the dimensions of the solar chromosphere, pro- 
vided that it contains the same gaseous materials. If, however, 
the surface temperature of the new star equals that of the photo- 
spheres of so-called ‘ white 5 stars, which, as we know, possess very 
extensive atmospheres, its gaseous molecules would tend to form 
an envelope of similar dimensions. How, I have shown in a 
paper in the Astronomische Nachrichten , Nos. 3822-3, that the 
extension of stellar atmospheres must be supposed to increase very 
rapidly if the surface temperature is raised. On the other hand, 
as shown in the same paper, our assumptions as to the tempera- 
ture of new stars immediately after the collision are practically 
