50 AVOYAGETO 



1776. Cape de Verde J/lands, by nwht or day, bem^ ivell afjurcd, with that 



November. -' . . *' " ' /• , . . 



'variation, that you are to the Eajlivard oj them. Another, in his 

 Chart *, lays down this variation ninety leagues to the 

 Weftward of them. Such a difagreement as this, is a ftrong 

 proof of the uncertainty of both. However, I have no 

 doubt, the former found here, as well as in other places, 

 the variation he mentions. But he fliould have confidered, 

 that at fea, nay even on land, the refults of the moil: accu- 

 rate obfervations will not always be the fame. Different 

 compaffes will give different variations ; and even the fame 

 compafs will differ from itfelf two degrees, without our be- 

 ing able to difcover, much lefs to remove, the caufe. 



Whoever imagines he can find the variation within a 

 degree, will very often fee himfelf much deceived. For, 

 befides the imperfe6lion which may be in the conftrudion 

 of the inflrument, or in the power of the needle, it is cer- 

 tain that the motion of the fhip, or attracftion of the iron- 

 work, or feme other caufe not yet difcovered, will frequently 

 occafion far greater errors than this. That the variation 

 may be found, with a fhare of accuracy more than fufficient 

 to determine the fliip's courfe, is allowed ; but that it can 

 be found fo exacftly as to fix the longitude within a degree, 

 or fixty miles, I abfolutely deny. 



* Mr. Dun. 



CHAP. 



