ASTRONOMY. 379 



The star a is the double star I 112G, the center being observed (the 

 distance is 1".3). The observations were reduced to 1867.0, the mean 

 proper motion of Procyon being' taken from Auwers' Fundamental 

 Catalogue. 



The results were — 

 From P— d: - = + 0".390 ± 0".055 <5a= 1".020 ± 0".140 



From P— b: - = + // .327 i 0".073 5a = I'M 01 =fc 0".250 



From P— c : - = + 0".307 ± 0".071 <5a = 1".007 ± 0".489 



From P— a ; - = + 0".3S3 ± 0".055 5a = 0".444 i 0."122 



where Sa is the correction to the adopted semi-diameter of the orbit, 

 0".9805. These results were found by comparing each observation of 

 Procyon with the single observations of all the comparison stars taken 

 on the same day, so that the four results are not independent of each 

 other. The author next treats the observations from 1863-G8, which 

 had been specially intended for a determination of the parallax, by 

 taking a mean of the comparison stars observed on one day and form- 

 ing the Aa between this mean and Procyon. In this way the relative 

 parallax of Procyon was found = + 0".299 ± 0".038. The star a is 

 Bradley 1107 and has been very frequently observed on the merid- 

 ian. From all available observations the proper motion was found = 

 — S .00255± S .00030, aud when this was inserted in the equations for 

 * aud da from P—a, - was found = + 0".398 ± 0".06t aud da = + 0".050 

 + OMOO, which results agree much better than with Auwers'. 



Since 1851 Otto Struve has observed the differences of E. A.»between 

 Procyon and the stars b and c, but the observations turn out to be far 

 less accurate than might be expected, so that the instrument cannot 

 have been firmly fixed in K. A. The observations of difference of dec- 

 lination from the same stars were specially intended to determine the 

 irregularity of the proper motion. Those from 1851-73 were compared 

 by Auwers with his circular elements and found to be very much bet- 

 ter represented under the supposition of irregular than of uniform 

 proper motion. The author shows that the entire series (1851-'82) 

 neither agrees with a uniform proper motion nor with Auwers' orbit. 

 The latter certainly gives smaller residuals, but they clearly indicate 

 corrections to the elements. As Procyon has only described about 

 three-fourths of a revolution since 1851, Auwers' period had to be 

 adopted. The following orbit represents the observations well — 



a = + 0". 0980 ;fc 0". 0354 



T = 1794. 9GG ± OK 440 



n = 9°. 02993 ± 0° 08072 (Auwers). 



The epoch agrees well with that found by Auwers, but the semidi- 



ameter is much smaller than Auwers' value, the difference being seven 



times greater than the probable error of either result. One of the two 



series of observations employed must therefore be affected by systematic 



