^6 Mr. Herschel's CcUaloguc 



other. I l^ave adopted thefe meafures on finding that I could 

 procure threads fine enough to fubtend only an angle of about 

 \" ix''-^ and that by this means there was no longer any 

 great difficulty of judging when the flars were centrally co- 

 vered by the threads. However, I do net know whether thefe 

 meafures, with flars at a confiderable diflance, may not bs 

 liable to an additional error of perhaps one fecond, owing to 

 the remaining uncertainty in judging of their exad central 

 pofition while the meafure is taking. 



The pofitions have ail been meafured (unlefs marked other- 

 wife) with a power of 460, adapted to an excellent microme- 

 ter, executed by Mefl'. Nairn e and Blunt, according to 

 the model given in the PhilofophicalTranfacSlions, vol. LXXI. 

 page 500. fig. IV.; but with a great and neceffary improve- 

 ment of making the wheel d^ d^ of that figure perform its 

 whole revolution ; by which means the two filk- worms- 

 threads may be adjufted to a greater degree of exa^lnefs ; for if 

 they are not placed fo as perfedlly to bife6l the circle, the two 

 threads will not coincide exactly after having performed one 

 femi- revolution, which they muft be made to do with the 

 utmofl rigour. I found the abfolute neceffity of this precau- 

 tion when I came critically to examine the pofitions of the 

 Georgium Sidus, as they are given in table III. Phil. Tranf. vol. 

 LXXI. p. 497. The meafures were afFe6led with a fmall and 

 pretty regular error, which I was at a lofs to account for ; and 

 the diflance oi this ftar being then totally unknown, I looked 

 for the caufe of the deviation at firft in a diurnal parallax of 

 that heavenly body ; but foon found it owing to the incon- 

 venience before-mentioned, of not being able experimentally 

 TO adjufl the moveable thread to that critical nicety which I 



have 



