for between 14° N. lat. and the Equator, in the ſpace contained between 23° and 26° W. long. the difference had not been more than 3°, or from 14° to 11°; whilſt in the ſame extent of ſouth latitude, between 26° and 30° W. long. the compaſs varied eight degrees to the eaſt, or from 11° to 3°. Might not the vicinity of the coaſt of Brazil be one of the principal cauſes of this difference?
The ſmalleſt variation obſerved was that of 1° 50′, in 25° S. lat. and 29° W. long. It cannot be doubted that the change of ſituation with reſpect to longitude has a much greater influence upon the variation of the needle, than change of latitude. The variation increaſed ſenſibly in proportion as we advanced farther eaſtward.
One of our officers who was taking the diſtances between the ſun and moon, in a very incommodious ſituation, with a copper ſextant made by Dollond, the radius of which was a foot in length, diſcovered a cauſe productive of error in the calculations, which one ſhould hardly have ſuſpected. The radii of this inſtrument, though very heavy, were ſtill ſo ſlender as to bend when it was preſſed with any degree of force againſt the breaſt, whereby the paralleliſm of the ſpecula was deranged. The ſame effect does not take place with ſextants made of wood, for their radii,
being
