THE PACIFIC OCEAN. 
321 
From this view of the time-keeper it appears, that, for 1779. 
near two years, it altered its rate very inconhderahly, and , Qaobe1, t 
therefore, that its error, according- to the Greenwich rate, 
if we had had no opportunities of corredting it, would have 
amounted only to il°. That afterward, at King George’s 
Sound, or Nootka, it was found to have varied exceedingly; 
of courfe, the longitude, by its Greenwich rate, was be¬ 
coming conliderably erroneous. About this time, it fhould 
be remarked, the thermometer was varying from 65° to 41 0 . 
The great eft alteration we ever obferved in the watch was, 
during the three weeks we were cruizing to the North; in 
which interval, it gave the longitude of the Eaft Cape with 
a difference of twenty-eight miles. I have marked the 
longitude of Saint Peter and Saint Paul, as given by the 
time-keeper, notwithflanding it flopped a few days before 
we arrived there; this I was enabled to do, from com¬ 
paring the longitude it gave the day before it flopped, with 
that given by Mr. Bayly’s watch, and allowing for the error 
of the latter. 
The ufe of fo accurate a meafure of time is fufficiently 
evident, from its furnifhing in itfelf the means of approx¬ 
imating to the longitude at fea, as may be feen in the above 
table. But, betides this, we were enabled, by the fame 
means, to give a degree of accuracy to the lunar obferva- 
tions, which they cannot otherwife pretend to ; and, at the 
fame time, by reducing a number of thofe obfervations to 
one time, obtain refults approaching flill nearer the truth. 
In furveying coafts, and afcertaining the true pofitions of 
capes and head-lands, it reaches the utmofl degree of prac¬ 
tical exadtnefs. On the other hand, it is to be obferved, that 
lunar obfervations, in their turn, are abfolutely neceflary, in 
Vol. III. T t order 
