METHODS OF OBSERVING THE LONGITUDE AT SEA, ^7 



voyages, yel if five, fix, or more, be taken to fea in the fame 

 Ihip, the longitude computed by each feparately, and the 

 mean of their refults taken, it will come exceedingly near the 

 truth, even at the end of three or four months. 



Another very great advantage refuUing from this method is, This method 

 that, (hould one of the time-keepers flop, it may be fet going ci^eof^ntgka"^ 

 again by the reft ; or fhould one of them be found to go very in winding up, 

 incorredly, its rate may be rejefted, and the longitude flill 

 determined by thofe which fhall be fuppofed to go well ; in 

 fliort, this method is almoft intirely free from all thofe accidents 

 to which a fingle time-keeper is liable. 



The two following examples, which are taken from the rates and affords a 

 of chronometers that have been determined in fixed obferva- r^°" JJ recifion 

 tories, will fliow the perfe6tion of this method more clearly. Examples. 

 In the firft example, which contains a period of 30 days, there 

 are only two inftances in which this method differs fo much as 

 a fingle fecond from mean time. Irtthe fecond example, con- 

 taining a period of fix weeks, the greateft error on any one day, 

 during that time, is only 5'\9, and the total error at the end of 

 that period, is only I'^.P. 



It may be faid that thefe are favourable fpecimens, but fup- If the chrono- 



pofing that thefe chronometers had gone ten times ivorfe than "^^^s ^^d per- 

 r f> o •' formed ten times 



they did, the greatefl error in the firft example would have worfe, the error 

 amounted to no more than li".3; and the greatefl error in the woi^'d have been 

 fecond example would not have exceeded a fingle minute. minute per 



month. 



F 2 FIRST 



