ee 
The Determination of Longitude. — 395 
achronograph. The stations must of course be connected 
by a telegraph line. The transitis the astronomical instru- 
ment used because by it time can be determined most 
accurately. This instrument should not be too large to be 
portable, as there are advantages in an observer constantly 
using the same instrument and each observer shouid 
therefore carry his instrument with him. The chrono- 
graph records the beats of the clock, the observations made 
by the observer at the transit instrument, and also affords 
the means of accurately comparing the clock at the distant 
station with his own clock. 
The usual programme of work is as follows:—-Telegrams 
are exchanged as to the probability of clear weather at the © 
stations. If the weather be fair the observers will make a 
determination of their clock errors by the observation of 
from 10 to 20 stars. This will occupy perhaps two hours, 
after which an exchange of time signals between the sta- 
tions takes place. If we call the stations A and B. Then 
A. sends signals which are recorded on its own chronograph 
and on the chronograph at B. On land line work it is suf- 
ficient to send for one or two minutes, or from 30 to 60 sig- 
nals and these may be direct clock signals, i. e., the beats of 
the clock at A sent along the line to B, or may be arbitrary 
hand signals coming in at irregular intervals on the chro- 
nograph record. B then sends an equal number of sig- 
nals to A, which are also recorded on the chronograph at 
B. For security against accident it is usual to send a 
second or check set from A to B and again from B to A. 
The object of sending signals in both directions is to elimi- 
nate what is called ‘wave time” or “transmission time” 
If, for example, we are exchanging between Montreal and 
Toronto, the Montreal clock being ahead of the Toronto 
clock on local time, the Montreal signals arriving at Toronto 
will be slightly retarded and the comparison of clocks on 
the Toronto Chronograph will givea quantity which will 
be two or three hundredths of a second too small. On the 
other hand the comparison on the Montreal chronograph 
will give a quantity two or three hundredths too large, the 
