ASTRONOMY: H. SHAPLEY 
133 
period variable stars. From one star to the next in all regions of the 
sky we observe continuous gradations of many constant characters, 
such as spectrum, intrinsic luminosity, and motion. That a star begins, 
passes through and ends its light-emitting career without change of 
spectral type is considered impossible. The universal and presumably 
justifiable assumption is that a given spectrum has developed by im- 
perceptible degrees from an earlier type, and will, given time enough, 
develop into a later one — that every star, in the course of its life history 
as a luminous object, passes through many if not most of the principal 
types and sub- types of the spectral series. This is indeed more than 
an ordinary assumption; it is a scientific creed. It is now generally 
accepted by astronomers as an inevitable fact, so nearly axiomatic 
that its demonstration is no more essential than the proof of the fact 
of organic evolution. The essential studies concern rather the methods 
of stellar evolution and its numerous details. 
That stellar evolution, as reflected in progressive changes of spectral 
type, has not been witnessed (the phenomena of novae are nebular 
rather than stellar) is, of course, no cause to doubt the basic assumption. 
Observations necessarily have been superficial, and time has been short. 
It is of interest, therefore, and probably of importance as a proof of 
stellar evolution to have observed a remarkable change in a star, not 
only of its light, as in the ordinary variable star, nor merely of its velo- 
city, as in the ordinary double star, but of its very physical make-up. 
And it is also of importance to know that, when the physical state of a 
star changes and its intrinsic brightness and surface temperature are 
affected, the spectrum varies (from one recognizable type to another 
through known intermediate stages) in exactly the same way as we 
presume the spectra of all stars have changed or eventually will change. 
It should be noted, however, that differences in spectra for the most 
part simply represent dissimilar conditions in the relatively thin atmos- 
pheric layer that is the source of the absorption lines; indirectly the 
deeper structure, even to the center of the star, may be and probably is 
involved. 
In the case under discussion in the present note the variation of 
spectrum is oscillatory in a period of a few hours, and does not represent, 
therefore, a secular, irreversible variation of spectral type. It does 
represent, however, an enormous change in the physical condition at the 
stellar surface, and exhibits the almost incredible feat of accompHshing 
in less than two hours time a change of spectrum that for the ordinary 
development of the ordinary star has been supposed to occupy unknown 
thousands or millions of years. 
