MOUNT WILSON SOLAR OBSERVATORY. 265 



positive and 15 are negative. Six stars show velocities of over 100 km., 

 the largest of these, that of Lalande 1966, being —319 km. Two stars of 

 type A show velocities exceeding 100 km. These are the first A-type 

 stars found to have very high constant velocities. Several pairs of 

 optically double stars are shown by their radial velocities almost cer- 

 tainly to be physically connected. 



Recent observations of the star 0. Arg. S. 14320, magnitude 9.2, 

 indicate that its motion in space must be extraordinarily rapid. The 

 star has a carefully determined parallax of +0''035 and a proper motion 

 of 3 ''76 annually. Its radial velocity as determined approximately 

 with the low-dispersion spectrograph is +290 km. This gives for its 

 motion in space 577 km. per second. There is a second star of the 

 same proper motion about 5' distant from it. The spectrum of the 

 star is G2. 



The number of spectroscopic binaries discovered during the year is 

 not large, the proportion of binaries among the solar-type stars which 

 have formed the principal part of the observing list being much less 

 than among the A- and B-type stars observed in previous years. A 

 peculiarly interesting binary is the parallax star Lalande 18397, magni- 

 tude 7.7. The spectra of both components are seen on the photographs 

 of this star, and both appear to be of type Kq. The maximum differ- 

 ence of velocity so far observed between the two components is about 

 120 km. 



Absorption of Light in Space. 



One of the principal effects of the absorption of light in space should 

 be a weakening of the violet light of the more distant stars relative to 

 the red light, when we compare such stars with those which are nearer 

 to us. The problem may be investigated in two ways: The fainter 

 and the brighter stars in a given portion of the sky may be photo- 

 graphed with and without a visual color screen, and the difference of 

 color between the bright and the faint stars may thus be determined; 

 or the spectra of stars of the same spectral type, some of which are 

 known to be near and others distant, may be compared with one 

 another. Each method has its disadvantages. The color-screen 

 method involves two assumptions : first, that the distribution of spectral 

 types is the same among the bright and among the faint stars ; second, 

 that the faint stars are faint on account of distance. The disadvan- 

 tages of the actual comparison of the spectra of the stars are that it can 

 be applied only to comparatively bright stars, and that the accumula- 

 tion of material is necessarily slow. 



The most accurate comparison of the spectra of stars of the same 

 type but of different distances may be made by obtaining their spectra 

 on the same photographic plate. In this way the effect of differences 

 of zenith distance and of variable transparency of the sky on the spectra 

 is nearly eliminated. It is also necessary to obtain practical identity 



