SOLAR ECLIPSE EXPEDITIONS— MITCHELL ^57 



and a single star on the other side of the Hne. As a consequence, the 

 Hght deflections to be determined depend to a high degree on the plate 

 ■constants used in the reductions. If, by means of least squares, the 

 scale correction is determined at the same time as the other constants 

 of the plate from the star observations themselves, that is, by follow- 

 ing the procedure adopted in 1919 and 1922, the 1929 plates give a 

 deflection of 1''75±0''13. Taking the values of all eclipses together 

 as solved by least squares, the observed value of the relativity de- 

 flection is l':79±0':06. 



The dift"erences between the various reductions of the 1929 eclipse 

 show the underlying difficulty of the whole problem, namely, the 

 practical impossibility of adequately separating the scale of the plates 

 from the star deflections. It, therefore, appears evident that work on 

 relativity displacements must be repeated at future eclipses by proc- 

 esses which will insure a far greater accuracy than has been attained 

 in the past. Such a recommendation is quite obvious — but unfor- 

 tunately it will be quite impossible to put tliis recommendation into 

 effect at any time within the next tw^o decades on account of the 

 poverty of stars in eclipse fields. 



Most astronomers and physicists are now satisfied that the theory 

 of relativity has been verified b}'^ observations made in three different 

 fields of investigation: (1) The motion of the perihelion of Mercury; 

 (2) the red shifts of lines in the solar spectrum; and (3) deflections on 

 eclipse photographs. In addition, Adams has measured a very large 

 relativity shift in the system of Sirius, and Trumpler has measured 

 red shifts in stars of type. Many new and interesting facts con- 

 cerning this theory will be accumulated in the next quarter of a 

 century of progress before there is another eclipse with a satisfac- 

 tory distribution of bright stars that will permit another thorough 

 test of the relativity displacement. 



