514 Proceedings of Royal Society of Edinburgh. [sess. 
same for all gases and vapours.* The identical behaviour of lines 
in the spectrum of Nova Persei pertaining to different chemical 
elements must be considered to contradict this explanation. For 
the same and also some other reasons we cannot perhaps accept 
a new theory advanced by Dr Ebert of Munich,! in which 
abnormal refraction is claimed as the principal cause of the 
peculiar duplex character of the Nova lines. Dr Ebert's con- 
siderations are based on the fact that in a medium, the spectrum 
of which shows distinct absorption maxima, the index of refrac- 
tion changes abruptly in the immediate neighbourhood of such a 
maximum, being greater on the less refrangible, and smaller on 
the more refrangible side. In his opinion, the light-radiation 
perceived in the bands of the Nova is not due to the radiative 
energy of the gas itself, but to light originally emanating from the 
incandescent surface of the star, which is abnormally refracted in 
the gaseous envelopes outside in such a manner that bright and 
dark bands are formed lying on the red and violet edges of 
the lines peculiar to the traversed gases. Three objections may be 
raised against this theory. Firstly, Dr Ebert’s theoretical intensity- 
curve of the bands, as we shall see, differs materially from that 
observed in Nova Persei. Secondly , his theory gives insufficient 
account of the presence of the bright bands after the continuous 
spectrum had disappeared. For obviously, if the continuous 
spectrum is, in Dr Ebert’s opinion, the conditio sine gua non for 
the bands, we are at a loss to explain how these bands could have 
possibly outlived, as they have actually done, the incandescence of 
the star’s surface. Thirdly , the effect of abnormal dispersion is 
by no means the same for all gaseous media. According to the 
electro-magnetic theory, it depends on the elastic resistance of 
the ions, a force which cannot be supposed to be the same for 
different atoms. According to observation, even lines of one and 
the same metallic vapour, e.g. sodium, behave quite differently. 
Hence the theory of abnormal dispersion seems to offer no 
explanation of that similarity in structure and character of the 
Nova bands, which, according to the observations, appears to be 
a fundamental feature of the Nova spectrum. 
* See H. Kayser, Handbuch der Spectroscopie, ii. p. 325. 
t Ueber die Spectren der neuen Sterne, A.N. 3917. 
