118 I. P. Kocu. 
As suggested on several occasions, the keeping of a fully satis- 
factory constant rate, especially as regards the observation watch, 
caused insurmountable difficulties. In order to illustrate this, I shall 
give below a summary of the rates of the observation watch on the 
days, when the moon observations were made for the determination 
of the longitude. This survey shows great oscillations, and this is 
particularly obvious in cases where two determinations of the clock 
rate were made on the same day (2x1 and 23/х1). From this it ap- 
pears that the observation watch, in spite of all the precautions 
taken, was not able to stand exposure to low temperature for any 
length of time. 
24. Sept. hourly rate -+ 05.16 
2 
RUE Sø 0.00 
18. Oct. = 5012 
1g, = =e - 0.04 
a =A + 0.32 
a = + 0.18 
DB = 22021 
a = + 0.33 
17. Nov ar + 0.01 
DD FE. = + 0.51 
=e = + 0.22 
D3.) = =: eee) 
er = + 0.40 
Be = 20.12 
5. = au ml 
Determinations of Longitude. 
For the determinations of longitude we used transits of the moon 
and an adjacent star, with the telescope unchanged in azimuthal 
direction. 
The principle of the determination is the following: By 
means of the time of the transit of the moon, of its declination and 
the latitude of the place in question, the azimuth of the moon is 
computed. The same method is used as regards the star, and the 
azimuthal direction of the telescope being unchanged one must — 
apart from errors of observation — arrive at the same azimuthal 
value in the case of both celestial bodies. It is, however, a presupposi- 
tion that the exact value of the right ascension of the moon is 
known, which in its turn presupposes a knowledge of the longitude 
of the place. The latter, which is the value looked for, is only 
known approximately; the right ascension and consequently the moon 
