OBSERVATIONS FOR TIME AND LONGITUDE. 163 
taken June 20th, and on June 24th showed 9h. 34m. 10s., when the same 
star had the same altitude, its daily rate would be 3* 6s. losing:— 
M. 
8. 
1 st time by watch .» .. 
9 
50 
8 
3 m. 55-91 sec. X 4 — 
— 
15 
4?*6 
Time watch should show 
9 
3 4 
24-4 
2nd time by watch.. 
9 
3 4 
10 
Losing in 4 days .. .. 
14-4 
This observation should only be taken when the star has a considerable 
altitude, so as to reduce the errors caused by refraction, and can only be 
used when a halt of some days is made, as any change in latitude would 
be followed by a change of altitude. 
Bate . 
It is but of little practical use to find the precise time of your observa¬ 
tion unless it is transferred to the watch. By taking the difference 
between the time resulting from the observations, and that shown by the 
W’atch, the error of the latter is found. 
The true time of any subsequent, or previous observation taken within 
a short time of the observation for time, can then be found by applying 
this known error to the watch time. 
If, however, the time is required some days later, it is necessary to 
know the rate of the watch, and this is obtained by repeating the observa¬ 
tion for time in the same spot after a few days, when the difference of the 
errors, divided by the time elapsed between the observations, will be the 
rate of the watch. 
H. M. S. 
Thu?, Error of Watch at Ojiji on 24th Sept., 8 A.sr., was 1 14 23 slow 
„ „ , 5 29th Sept., 8 a.m., was I 15 17 ,, 
Difference 5) 54 
Rate of Watch = io-8 losing 
Then, supposing that observations for longitude, say, by occultations, 
Were obtained on the 26th without being able to obtain observations for 
time on the same day, the time can be found by applying the rate to the 
previous error, thus• 
m 2 
