26 M. Bessel on the Declination of' the Jixed Stars. 



the course of my own practice; and therefore, would have 

 laid no particular weight on the declinations by my own obser- 

 vations, if I had not believed that I had contributed somewhat 

 to their certainty by previous proofs of my circle. My confi- 

 dence was increased by the simplicity of the instrument, which 

 allows the observer to become perfectly acquainted with every 

 single step: added to which, his being able to compare his 

 observations on both sides of the circle, affords him a proof 

 of which he is entirely deprived by the use of the repeating 

 circle. Even after the doubt expressed against the correctness 

 of my determinations, I was unable, by the most rigid examina- 

 tion of the instrument, and method of observation, to discover 

 a constant error of more than one second : and even this limit 

 I should consider too great, if I had completed my earlier 

 purpose of determining the particular errors of division for 

 each of the fundamental stars, by my method of repetition 

 {Wiederholimgs-Methode), but which I have been prevented 

 irom doing, not only by other occupations, but also from the 

 expectation of shortly receiving the meridian circle of Reichen- 

 bach. From this there was much accuracy to be expected in de- 

 tail, whereas I concluded that in Gary's circle I might venture to 

 confine myself to such rules as aimed at reaching certainty in 

 general. I considered therefore that time would solve this 

 difficulty, and by progressive improvement add to the science, 

 pf observation. 



This expectation is greater from the meridian circle, in 

 the completion of which Reichenbach has crowned his im- 

 mortal labours. Great, however, as the perfection of this in- 

 strument is, I do not think that the astronomer can venture 

 to disjDense with the trouble of personal investigation ; taking 

 for granted that he possesses the means of investigating it with 

 that accuracy which is commensurate with the great perfec- 

 tion of the instrument. According to my idea, that instru- 

 ment must be corrected [eliminirt'] by all the observations ; 

 and only when these can be shown can the results be consi- 

 dered incontrovertible. I have uniformly kept this object in 

 view ; but whether I have thereby arrived neaver the truth, 

 remains to be proved. But this I will venture to say, that I 

 have not left unexplored any source of error, the possibility of 

 which I could conten)plate. From these proofs it is now ap-' 

 parent that the plan of the circles requires only two cor- 

 rections ; viz. in regard to the error of the divisions, and the 

 bending of the telescope. Still, if we hope to gain as correct 

 determinations by the meridian circle as by Bradley's muraj 

 quadrant, it would be necessary to correct even these, by a 

 i^ew fundamental determination of the refraction : on which 



account 



