ON AN IMPROVED REFLECTING CIRCL.E. 5 



Mr. Mayer propofed to complete the limb of the fextant, i. He made the 

 making a whole circle, with the horizon glafs moveable round ^ d ^ n A^Ta. 

 the centre, with an additional index, which I fhall call the he fixed the ho- 

 horizon index, in order to diftinguifh it from the centre index, to **?£* u P on 



° an index move- 



which the centre glafs is attached. This inftrument is repre- able on the fame 

 fented in Plate I. Fig. 1; and the manner of ufing it is centre, but in- 



• i • • • „ dependant or 



as follows : after the index A is let at o, (the beginning of that which car- 

 the divifions,) the two glafTes are rendered parallel, as is ne * t}ie &*{* 

 ufually praclifed with Hadley's quadrant, by moving the hori- gj a fs. 

 zon index B, till the horizon of the lea, (or the fun, or any ufe. Adjuft the 

 other object,) or its direct image, and the doubly reflected j» !afl " parallel 

 image of the fame, feen through the telefcope, coincide. i n j e x< 

 After fixing the horizon index in that pofition, the centre in- 

 dex A is to be moved, in order to meafure the diftance of the Obferve the an- 

 two obieds S and L, (which I fhall fuppofe the fun and moon,) f^^ £? 

 by bringing into contact the doubly reflected image of the fun 

 with the direct image of the moon, feen through the telefcope. 

 The centre image will then be at M, and the arch o M might 

 give, as in the fextant, the angular diftance required ; but the 

 eonftruction of the circle renders it eafy, in this pofition, to Make the glaflea 

 efTed again the parallelifm of the glafTes, and to make another "gh ^ ei by 

 obfervation of the contact, in the like manner as from o ; index. 

 which operation will bring the centre index to N. The index Obferve the an- 

 will then give o N, or double the diftance ; and, as it muit g e as e ore * 

 be divided by 2, in order to have the angle required, the er- Adjuft for parai- 

 rors of divifion and eccentricity, which, together, I fhall call ,cWm ? an ^/ c ' 

 the error of the injirument, will be Iikewife reduced to one half, vation. 

 It is obvious, that by fucceflive repetitions of tbe fame procefs, And as this pro- 

 triple, quadruple, &:c. the diftance may be obtained, and the cefs . ma >'. bc 

 faid error further reduced, in the inverfe ratio of the multi-pi ea fure, and 



plication of the diftance, to any degree of approximation re- the firft ' nde * 



• , J to rr will pafs over the 



<l llired - meafure of the 



The method of rendering the glafTes parallel, by means of angle as many 

 the horizon of the fea, is not accurate, on account of the in- af^"^^^ . 

 diftinelnefs of the images; and, when the fun is ufed for that Divide the whole 

 purpofe, the obfervation becomes fatiguing to the eye. The Jj^ b by l i e 

 repetition of that obfervation, by one or the other method, tition», and the 

 remained therefore a confiderable inconvenience attached to quotient will oe 

 Air. Mayer's circle. The author himfelf feem,s to have been g i e j t h e error* 

 of that opinion, as he propofed to provide the inftrument with of the > nftru - 

 a diagonal rule, fixed upon one of the indexes, fo that the™™^""* tl JJ 



Other f*me ratio. 



