222 CARNEGIE INSTITUTION OF WASHINGTON. 



The color index is given by the differences between the photographic 

 and the photovisual magnitudes. 



The measures and reductions already completed indicate that the 

 method should give satisfactory determinations of color differences, 

 which will serve for the study of the problem of absorption of light 

 in space. Professor Hertzsprung's results in the same field, obtained 

 with the aid of a large objective grating, are described in the follow- 

 ing section. 



Professor Hertzsprung's Investigations. 



In the last annual report an account was given of the observations 

 undertaken by Professor Ejnar Hertzsprung, of the Astrophysical 

 Observatory of Potsdam, who spent four months on Mount Wilson for 

 the purpose of determining the effective wave-lengths of the light of 

 faint stars. Since his return to Europe Professor Hertzsprung has 

 measured and completely reduced the plates he obtained last year 

 with the 60-inch reflector for the region containing the cluster 

 N. G. C. 1647. The remaining regions will be investigated as soon 

 as a suitable measuring-machine becomes available. 



Twelve plates of N. G. C. 1647 were measured in Professor Kap- 

 teyn's laboratory during a visit of Professor Hertzsprung to Groningen. 

 Half of them are Lumiere "Sigma," the other half Seed "27" plates. 

 The times of exposure range from 6 seconds to 30 minutes, in order 

 that the same star may show images of widely differing intensities, 

 thus enabling the observer to reduce all the measured wave-lengths 

 to the same intensity of image. 



For obvious reasons the effective wave-lengths are found to be 

 much more precise for the brighter stars than for the fainter ones. 

 Over 200 stars were measured. For the 50 brightest the mean error 

 is on the average about ±8 a, corresponding to ±0.04 magnitude 

 in the difference between photographic and visual magnitudes (color 

 index) . The resulting effective wave-lengths show a very systematic 

 change with apparent magnitude, in the direction indicating that the 

 fainter stars are redder than the brighter ones. The Lumiere "Sigma" 

 and Seed "27" plates agree closely in making the change about 20 a for 

 each magnitude. There is only a single faint star (mag. 12.4) which 

 appears decidedly white. 



The phenomenon may be due in part to the fact that the bulk of 

 the stars belong to a single physical group. For such a group analogy 

 with other known cases would lead us to expect increasing redness 

 with decreasing absolute brightness. On the other hand, many of 

 the stars can not belong to such an assumed group. We thus have a 

 strong indication either that in the region considered there is but one 

 star which is certainly of the first type or that the color of the stars 

 is due to selective absorption in space (?). 



