OF ARTS AND SCIENCES. 



347 



Where these proper motions are sensibly erroneous, usually too 

 large, the true fractions of the year were employed ; so that in no case 

 will sensible error arise from emjjloying the fractions to tenths only. 



It is well known that the right ascensions of polar stars as deter- 

 mined by different astronomers seem to be affected by a different per- 

 sonal equation from equatorial stars. Thus, for instance, the stars 

 determined by M. Gonnessiat, at Lyons, for 1883.0 (Comptes Rendus, 

 for August 13, 1883) compare as follows with the right ascensions 

 of the Berlin Jahrbuch and those of the present paper ; G. denoting 

 the Lyons observations, B. the values given by the Jahrbuch, and 8. 

 those here given. 



The tendency to the positive sign, both in B. — G. and S. — ' G. is 

 here uiuTiistakable. If we assume (the propriety of this assumption 

 might perhaps be questioned) that these values are mainly px'opor- 

 tional to the secant of the declination, we shall have, — 



* From the " Annales du Bureau des Longitudes." 



