P LA PEROUSE. $f 



grec to the ?oth, the variation of the com- 

 pafs extended from n to 3 towards the weft, 

 which makes S of yaria ion fouth of the equator, 

 while north of it the needle had varied only 3 in, 

 a fpace almoft as great. 



Is not the proximity of the coaft of Brazil one 

 of the principal caufes of this difference ? 



The lean variation obferved was i 50' in the 

 latitude of 25 fouth, and longitude of 29 weft. 

 There is no doubt that the difference of longi- 

 tude in thefe fcas has much more influence on 

 the variation of the magnetic needle, than the 

 difference of latitude. The variation fenfibly in-, 

 creafed as we advanced to the eaftward. 



One of the officers took the diftances of the 

 moon from the fim, in a pofition Ibmewhat 

 inconvenient, with one of Dollond's brafs fex- 

 tants, the radius of which was two double deci- 

 meters : he perceived a caufe of error, which it 

 would have been difficult to fufpect. The radii 

 of this inftrument, although extremely heavy, 

 were ftill fo weak as to bend, and thus cauie 

 a derangement in the pararellifm of the glaffes, 

 when the obferver refted it with a little weight 

 ^gainft his breaft. The fame effect: did not take 

 place in the wooden fextants, becaufe their radii 

 being much thicker, do not yield to the force 

 which affects thofe made of brafs. . 



This fource of error is an additional reafon for 

 giving an exclusive preference to the reflecting 



circle 



