264 CARNEGIE INSTITUTION OF WASHINGTON. 



The immediate importance of the phenomenon lies in its applica- 

 bility to a determination of the distances of the cloud-forms of the 

 Milky Way. When the correlation is clearly defined, the method 

 should give excellent results, for the relation between mean color and 

 luminosity is well established by the spectroscopic parallaxes of the 

 Cepheids and pseudo-Cepheids, and especially by the extensive 

 determination of colors and absolute magnitudes in globular clusters 

 by Mr. Shapley. Even when the relation is more or less completely 

 obscured, valuable indications of distance — minimum distances, at 

 least — should be possible. A survey of the evidence bearing on corre- 

 lations of luminosity with spectral type and color (p. 275) indicates 

 that the blue stars in the galactic clouds are approximately of zero 

 absolute magnitude. The adoption of this value gives at once an 

 approximation for the distance, which can only be increased by the 

 assumption that the clouds, like the loose aggregations of B stars in 

 our own system, also contain blue stars of much higher luminosity. 



Very provisional results for four regions are: 



Selected Area 88 14000 parsecs. 



Scutum Sobieski 7000 



4h 2m -f 33° 15' Perseus 8000 



4 16 +26 3 Taurus 7000 



At a minimum, therefore, the distances of the blue stars in these 

 fields, of apparent magnitudes 14 to 15.5, seem to be of the order of 

 20,000 to 50,000 light-years. 



The obscured areas for which results are complete are in Perseus 

 and Taurus. The stars are predominantly red, but thus far no evi- 

 dence of an excess of color over that to be inferred from spectral type 

 has been found which can certainly be attributed to the obscuring 

 material. Mr. Hubble has made long exposures of the regions in ques- 

 tion with objective prisms on the 10-inch refractor. As far as the 

 spectral classification can be pushed, it indicates an abnormally high 

 percentage of late types, many of which, according to Mr. Lindblad, 

 are certainly dwarfs, agreeing closely with the colors measured by 

 exposure ratios and checked by polar comparisons. The inference 

 is that the obscuring material is comparatively near and obstructs, with- 

 out perceptible scattering, the light of the more distant stars. On this 

 hypothesis, the stars seen within the boundaries of these areas would 

 lie between the observer and the obscuring cloud and include a large 

 percentage of dwarfs, which in the nature of the case are of high color. 

 Thus in one of the fields in Taurus, at 4*^ 15"", -|-28° 10', the types of 

 7 well-determined stars within the boundaries of obscuration (m = 8.5 

 to 12.2) are all G or later; 6 of the 7 stars are dwarfs. The brightest 

 is a KO giant. The measures show a small color excess, less in amount, 

 however, than the uncertainty affecting it. Assigning to the dwarfs 

 mean absolute magnitudes corresponding to their types, we have 5.5 



