196 CARNEGIE INSTITUTION OF WASHINGTON. 



of 20 kilometers for the solar velocity. The material was revised, 

 assuming the following rectangular coordinates of solar motion : 



x= -2.3 km. 2/ =-13.2 km. 2= +14.9 km. 

 corresponding to a solar apex A = 260?l, D = +48?0, and to a solar 

 velocity of 20.04 km. The reason for this choice appears later. As a 

 result of the new elements of solar motion the distribution of stellar 

 apices changes as follows: 



Group I. Rejected 4, added 2 Group II. Rejected 6, added 6 

 with a material correction 



d7r = 0r057+0.005dp' 



and a maximum value for dir of 0''082. 



Solving the conditional equations of the two groups based on the 

 new solar motion, we obtain the following coordinates of the two planes : 



Plane I (30 apices) 7i = 189?47=t3?98 ki = +34?84±2?62 



Plane II (31 apices) 72 = 144.63±6.51 K2= -50.04±4.06 

 Comparing these results with those first given above leads to the impor- 

 tant conclusion that the two solutions agree within a few degrees. The 

 mean errors are of the same order and the deviation from perpendicu- 

 larity is slight (-0.0125). 



The second important point concerns the position angle. This 

 angle is independent of the radial velocity. Moreover, the partial 

 derivatives with respect to the other elements are very simple. These 

 important facts lead to a determination of the distances for all apices 

 certainly belonging to the velocity planes found above. The further 

 investigations may be discussed in two ways: 

 (1) using the observed radial velocities we are 

 able to determine the corresponding solar mo- 

 tion. A preliminary computation results in the 

 second value of solar motion used above. It is 

 interesting to see that this result is in good 

 agreement with the position of the solar apex 

 derived by L. Boss from G-type stars, although this type is not pre- 

 dominant in the star material here treated, as may be seen from the 

 accompanying table of distribution of types in the two groups. Then 

 (2), using a given solar motion, we are able to compute the result- 

 ing radial velocities and to compare them with the observations. In 

 this way we may obtain some insight into solar parallax and the prin- 

 ciple of relativity, if we study the measured displacements of spectral 

 lines. 



In these and similar investigations, where we have to deal only with 

 the possible influence of variations with respect to the principal ele- 

 ments, the results obtained are always a function of the assumed 

 weights and we have to pay special attention to this delicate subject. 



