ASTRONOMY: W. S. ADAMS 
359 
paratively narrow single absorption lines unaccompanied by emission bands 
cannot be explained in this way. Their character, as well as the law of varia- 
tion of displacement with wave-length, also precludes the agency of pressure 
(assuming that negative displacements ^can be produced by pressure) and the 
Zeeman effect. No dependence of anomalous dispersion upon wave-length is 
known, even if it were adequate to produce lines of such a character with such 
displacements. 
In some respects the Doppler effect accounts most nearly for the results 
observed. Motion in the line of sight would produce displacements directly 
proportional to wave-length and might leave the spectral lines well-defined. 
The velocities involved, — 750 km. in the case of Nova Geminorum and Nova 
Aurigae, and — 1500 km. for Nova Aquilae and Nova Persei, though large, 
are of the order found for some spiral nebulae, and not many times greater 
than those observed in some of the solar prominences. If the more refrangible 
component of the hydrogen, helium and calcium lines is considered, the veloci- 
ties amount to —1500 and —2200 km. respectively. These values are for 
motion in the line of sight. If a generalized form of the Doppler principle is 
used a part or the whole of the effect might be referred to a rapid change in 
the thickness of stratum of gas producing the absorption, or in its refractive 
index. This hypothesis was suggested by W. Michelson 4 to account for the 
high velocities observed in solar prominences, and by Paddock 5 in the case 
of Nova Aquilae. Some objections to this view are the rate of change in re- 
fractive index or in thickness of the gas which would be required, especially 
since a decrease of index would be needed to produce displacements toward 
shorter wave-lengths; the harmonic relationship found among the displace- 
ments; and the relatively narrow character of the absorption lines. 
The last two objections might also apply to motion in the line of sight. The 
suggestion, however, may be made that the absorption lines are produced in 
a shell of gas which is moving radially outward from the star with a high 
velocity. If the size of this shell is large as compared with that of the body 
of the star it is evident that an area of the shell only equal to that of the star 
would be seen in projection against the latter and would give absorption lines, 
and that all of the gas within this area would have large components of ve- 
locity toward the observer. This would result in comparatively narrow ab- 
sorption lines. The remainder of the shell of gas would give an emission spec- 
trum, and the combination of the widely different velocities would result in 
very broad bright bands with their centers nearly undisplaced. This is in 
accordance with observations. With these high velocities the interval of two 
days between the outburst of the star and the appearance of the prominent 
absorption spectrum would be sufficient for the gas to reach a great distance 
from the surface of the star. The hypothesis would, however, leave unex- 
plained the apparent acceleration of motion during the period of observation 
of the absorption spectrum; and the nearly constant character of the emission 
bands after the disappearance -of the absorption spectrum would point rather 
to their origin in the star itself. 
