STELLAR ASTRONOMY 417 



denote: mean x of the proper motions to the right, resp. to the 

 left of the great circles through the Apex. 



If there were real symmetry, the two ought to show insignificant 

 differences. 



The following table shows the actual value of the differences. Mean 

 values were computed for different galactic latitudes by combining 

 the results of regions at equal distance from the Milky Way. 



Mean value of X R X L (centennial motions) . 



\ApexA 2736 291 276 



Gal.lat.\ D +29 +34 +19 



-40 to -90 + 4".2 + 3".6 +3".2 



-20 to -39 +2 .2 +1 .6 +2 .4 



to -19 + 1 .8 +1 .2 +1 .4 



+ 1 to +20 .0 -1 .0 +0 .3 



+21 to +40 -2 .4 -2 .5 -1 .7 



+41 to +90 -4 .8 -3 .8 -4 .9 



Mean absolute value of =3*7. 



This table has been separately derived for stars of Secchi's first 

 and second type, and for those whose spectrum has not yet been 

 determined. The result has been practically the same for all. Also 

 the other component of the proper motions, at right angles to the 

 former, has been investigated. Divergences are shown of a similar 

 character. Finally in the distribution of the numbers of proper 

 motions over the four quadrants, the phenomenon is as evident as 

 it is in our table of the x's. About its reality there thus cannot be the 

 slightest doubt. 



To what is it to be attributed? The amount of the divergences 

 summarized in Fig. 3 and in our table is so enormous that an ex- 

 planation of them by an uncertainty of the precession or about the 

 systematic corrections required by Bradley's observations is at once 

 excluded. Besides, for both these elements, the best available values 

 have been used. 



More probable seems an error in the adopted position of the Apex. 

 A glance at our Fig. 3 may even perhaps seem to favor such a view, 

 though a closer examination will show that by displacing this point 

 we shall certainly not succeed in making the phenomenon disappear. 

 To show this more convincingly I have repeated the calculations on 

 which our last table rests for two other positions of the Apex, one 

 differing widely in Right-Ascension, the other in Declination. The 

 results are shown in the table. They are practically the same as before. 



Or have we to do with a common motion of the whole of the stars 

 which have contributed to any one of our figures? Such an explan- 



