MOUNT WILSON SOLAR OBSERVATORY. 267 



the dense central regions of clusters. Incidentally a number of plates 

 have been made for the purpose of studying the variable stars in 

 globular clusters. 



Computations of the Orbits of Eclipsing Binaries. 



Several investigations relating to eclipsing binaries have been under- 

 taken by Mr. Shapley, chief among which are determinations of spec- 

 troscopic and photometric orbits for RX HercuUs and of photometric 

 orbits for Algol. For the former it was possible to obtain the spectro- 

 scopic orbits of both components from only 3 plates, as the light curve 

 had already furnished data for a number of the orbital elements. The 

 derivation of the orbit of Algol involved the first application of the 

 hypothesis of intermediate degrees of darkening at the limb. The 

 computations also considered the possible effect on the light-curve of 

 the third body in the system. 



Other computational work includes the formulation of simple rela- 

 tions for the upper and lower limits of the mean density of echpsing 

 variables, the discussion of low densities among second-type stars, and 

 the revision of certain double-star orbits. 



Cepheid Variables. 



Observations on faint variable stars by Mr. Shapley with the 60-inch 

 reflector have included the investigation of the hght and color changes 

 of a few stars of the "cluster type." For XX Cygni, which is the 

 shortest-period variable on record, the observations and reductions 

 have been completed. The main result is the proof from Mount Wil- 

 son and other observations of the changing shape of the light curve. 

 The lack of regularity in the curve, together with a number of other 

 considerations based on peculiarities of the light and velocity varia- 

 tions of Cepheid and ''cluster-type" variables, leads to the conviction 

 that the long-standing hypothesis that Cepheids are double stars is 

 insufficient. In fact, the evidence now in hand shows the impossibility 

 of any theory so far proposed with that assumption for its basis. 



STELLAR SPECTROSCOPY. 



The results of an investigation by Professor Kapteyn and Mr. 

 Adams, completed in the autumn of 1914, have enlarged to a consid- 

 erable extent the problem of radial velocity observations with the 

 Cassegrain spectrograph. The main feature of this research, to which 

 reference was made in the last annual report, was the proof of the 

 existence of the two star-streams among the very distant stars of the 

 solar type of spectrum. In addition, however, two other results were 

 derived : first, a relationship between radial velocity and proper motion 

 for the stars of the F, G, K, and M types of spectrum; second, a certain 

 amount of evidence tending to indicate, in the case of the K stars at 

 least, an increase in radial velocity with decrease in brightness. The 



