ASTRONOMY: SHAPLEY AND SHAPLEY 
453 
motion. But the many serious obstacles in the way of the double star 
hypothesis have been pointed out in a previous paper/ and the sug- 
gestion made that the observed facts may accord much better with 
the assumption that the hght and velocity variations are both due to 
disturbances on a large scale of the radiating surfaces of single stellar 
bodies. Such periodic disturbances, which have in sun-spots a not 
very distant analogue both as to spectral displacements and light vari- 
ations, would have as an underlying cause the long-enduring free oscil- 
lations that may be set up in a gaseous mass in a variety of ways, such 
as the collision with or close approach to another body, the perturbations 
by an obscure companion, or the adjustment of mass or temperature 
dissymmetries. 
Among the arguments against the double star interpretation of Ceph- 
eids is their property of definitely changing color (therefore, of chang- 
ing surface temperature) with the variation in the brightness of the 
star, thus indicating that the phenomenon is peculiar to the radiating 
surface and not merely geometrical as with the eclipsing double stars. 
Another point is the lack of regularity in the time of the principal phases 
and in the nature of the light variations. It is in connection with these 
two arguments that the present investigation hopes to contribute to the 
explanation of the Cepheid variables. In particular the present dis- 
cussion of the light variations of the star XX Cygni challenges the 
frequently iterated assertion that the short period variable, because of 
its supposedly great precision and regularity of performance, is a veri- 
able timepiece. In fact the suggestion is on record that the unit of time 
might possibly be based on the periodic return to maximum brightness 
of certain short period variables rather than on the rotation of the earth. 
While in most cases so far as now known the mean periods of such vari- 
ables have, to be sure, but little or no secular or periodic variation, the 
study of XX Cygni, in agreement with preliminary results for similar 
variables, has shown definitely that the light curves change rapidly and, 
it may be, erratically both in time of maximum and in character of vari- 
ation. The details of this work will appear in the Astrophysical Jour- 
nal;"^ some of the chief results are outhned below. 
In a monograph on the light variations of XX Cygni, Kron gives 
for the visual range 0.76 mag., based on the observations of seven 
observers, and for the photographic range, from observations by Park- 
hurst and Jordan, 0.63 mag.=^ This result is unusual in that the visual 
variation is the larger; every other variable star for which we have 
such information has a greater range photographically than visually. 
The difference between the ranges is a direct measure of the change in 
