Now there remains to be decided what weight must be given to 
each of the 6 results. Although it seemed at first that for each watch 
we should have to adopt a different aceuracy peculiar to it, it 
appeared after all very difficult to determine this intrinsic accuracy. 
So e.g. for N°. 84 we should have had to adopt a rather high weight 
according to the observations at Jidda and yet it went very irregu- 
larly during the journey to Mecca. So ultimately I adopted the same 
intrinsic accuracy for each of the watches; nor could greater weight 
be given to the chronometer of Dent, as a travelling instrument, 
than to Lerov’s watches. 
Hence I have given weights to the 6 results proportional to the 
number of journeys in which each watch had been used, and besides 
only a weight of 0.5 to watch 77 owing to the discontinuity during 
the stay at Mecca. 
So 1 obtained: 
DIFFERENCE OF LONGITUDE JIDDA—Mecca. 
Dent 2527 + 2™ 54562 Weight 3 
WarcH 7 34.64 i 2 
Ee Li Ca ee x 0.5 
- 80 55.15 se 2 
a 81 35.58 en i 
84 33.70 pe 1 
From the agreement of these six values inter se there follows as 
mean error for weight unity + 0.66 and the final result from the 
6 watches is found to be 
Difference of longitude + 273474 + 0°22. 
It is clear from the foregoing that a derivation of the difference 
of longitude from partial results for each journey must lead to a 
less advantageous combination of the observations. Yet I want to 
show that a final result obtained in this way does not differ much 
from the above given. 
We then obtain, indicating the journeys Ji—Me.—Ji. and Me.— 
JiimMe. respectively by M. and J. 
M I M II M Ill J I Jil 
Dent 36°40 55°30 93:28 35:89 33:57 
7 34.82 36.20 33.58 
id 3+.56(4) 
SO 35.37 34.93 
81 35.58 
84 33.70 
Mean 35°78 „4-96 3453 39°89 33:58 
