398 Canadian Record of Science. 
(equal to half the difference of the results) and gives a 
mean value for the difference of longitude of 9™. 475.516. 
In the final reduction of such work, a process known as 
“ weighting” is employed. That is, some portions of the 
work are given a greater value than others, depending up- 
on the accuracy and the number of the observations obtained. 
This does not, however, and should not in good work 
materially affect the result. Thus, in the work referred 
to, the mean difference after weighting is 9™. 47°.414 or 
practically the same as the arithmetical mean. 
Another method of conducting a longitude determination 
telegraphically, is to select certain stars which are to be 
observed at both stations and recorded on both chrono- 
graphs. Thus the observer at the easterly station observes 
a star and records its passage over the wires of his instru- 
ment on the chronograph at the western station as well as 
on his ownchronograph. After the lapse of the necessary 
time (the difference in longitude) the star comes to the 
meridian of the western station when the observer there, 
also records it on both chronographs. This method has 
the practical disadvantage that it requires the use of a tele- 
graph line between the stations during the whole evening 
_ instead of during a few minutes as in the method usually 
employed. It has the advantage, however, of being free 
from errors due to star places (the time given in the cata- 
logue for the star’s passage of the meridian). ‘This small 
source of error can also be eliminated in the other and usual 
method if the precaution be taken to observe the same 
stars at both stations. In the longitude work referred to, 
five stars were observed on June 23rd., at both Montreal 
and Cambridge, and recorded on the Montreal chronograph 
giving a mean difference after correction of 9". 47°.406 ; 
three stars were observed at Montreal and Cambridge 
which were recorded on the Cambridge chronograph, giv- 
,ing a mean of 9™. 478.523. The mean value for the differ- 
ence of longitude from both sets is therefore 9". 47°.465, a 
result very slightly less than that obtained by the main de- 
termination. This result was given some weight in com- 
