246 PROCEEDINGS OF THE AMERICAN ACADEMY" 



his time for twelve years to the observation of this single star ; the 

 result of this was the position for 1820,* 



0"^ 57" P.505 +88° 20' 54".27, 



which, it will be seen, agrees nearly with observations of other astron- 

 omers. For the proper motion he compared this with Bradley's posi- 

 tion (Bessel's own second^ reduction), and from these two determined 

 the A. R. and Dec. for every year from 1750 to 1850. 



Like all Bessel's works, this has fully stood the test of time ; the 

 correction which now appears necessary can only be deduced by 

 comparison with a great number of observations. 



The other computations of the position and proper motion of Pola- 

 ris to which allusion has been made are Wolfers's in the Tabulte Re- 

 ductionum and Dr. Gould's. The first agrees very closely with Bes- 

 sel's. It is derived from a slightly different and more accurate reduc- 

 tion of Bradley's observations (the one employed in Table I., and 

 cited in the notes), and from the mean of the best observations made 

 about 1830. It does not, however, agree quite so well with observa- 

 tion as Bessel's. 



The other one (Dr. Gould's) is also less accurate than Bessel's ; 

 both in A. R., for 1855.0, and in proper motion ; in the former ele- 

 ment its error is about double that of Bessel's, in the latter rather 

 more. 



The investigation (Astr. Journal, VoL VI.) seems hardly com- 

 plete. We miss in it Bessel's, Argelander's, Carhni's, and Henderson's 

 observations; Airy's Cambridge determination, and, of those made 

 under Struve's direction, the latter half of the Dorpat series (made by 

 Preuss) and Schweizer's Pulcova observations ; besides the long series 

 made at Paris under Arago, and several others of less authority. 



We must also regret that inaccuracies have crept in. Bradley's 

 place for 1755 is wrongly reduced (by about 2' of time) to 1855, and 

 again in another place to 1850 ; to Piazzi's for 1800, has been added 

 about 3*.6, which seems to be the "probable uncertainty" (error veri- 

 similis)' of a difference between Argelander and Piazzi. Gi'oom- 

 bridge's for 1810 has been changed by 4'.3 on account (it is stated) 

 of the error of Groombridge's meridian mark. But this error had 



* It will be seen in Table I. that I have given separately the results for 1817 and 

 1824, of the two transit instrumQnts which Bessel then used, 

 t Using as for his own observations Lindenau's nutation. 



# 



