Ch. I. SOUTH AMERICA. 17 



re6ted in 1744, from other oblervations and journals 

 by MefT. Montaine and Dodfon of London, I jftiall add 

 fome reflexions, in order to expofe the negligence in 

 conftrudting the magnetic needles. 1. It appears that 

 the variations obferved by Don George Juan do not 

 agree with mine, which is not to be attributed to a de- 

 fed: in the obfervations. This is fufficiently evident 

 from comparing them. The differences between thofe 

 obferved by Don George and thofe on the chart, are 

 nearly every-where uniform the moil confiderable be- 

 ing a degree and 30 minutes-, one making the variation 

 2° 30', and the other a degree only. This probably arofe 

 from the motion of the fhip, which hinders the needle 

 from being entirely at reft or from the difk of the fun, 

 by reafon of intervening vapours, not being accurately 

 determined, or fome other unavoidable accident; the 

 error, when the difference is lefs than a degree, being 

 fcarce perceivable in thefe obfervations. Thus, on a 

 medium, the rational concluHon is, that the needle ufed 

 in thefe obfervations varied a degree and forty minutes 

 lefs than thofe when the map was conftru6led. 



The fame uniformity appears in the differences be- 

 tween my obfervations and the chart ; but it muff be 

 obferved, that having ufed two different needles, the 

 particulars of each nearly correfpond, fo that between 

 the five firft, the greateft difference is of 40 minutes, 

 which intervene between the fmallcff difference of 2° 50', 

 and the greateft of 3° 30 . Hence, taking the medium 

 between both, the difference between my obfervations 

 and the chart will be 3^ 16', the latter being fo much 

 lefs than the former. The three laft do not want this 

 operation, the difference of 1° 30' being equal in all, 

 and the variations refulting from thefe obfervations are 

 alfo lefs than thofe delineated on the map, the variation 

 having paffed to a different fpecies , namely, from N. 

 W. to N. E. This demonftrates, that the firft needle 

 I made ufe of, whether it had been ill touched or the fteel 

 not accurately placed, varied 1° qo wefterly lefs than 



Vol. I. C ^ thai' 



