420 Trajisactions of the Boyal Society of South Africa. 
I pass on now to the subject which I have selected on which to address 
you to-night, and which I propose to entitle — 
" Some Eecent Improvements in Transit Observing." 
Whenever I happen to be in Cape Town about midday I notice a 
certain instant when almost every one, who is the happy possessor of one, 
proceeds to take his watch from his pocket and ascertain its indications. 
He is prompted to do so by the firing of the gun on Signal Hill, and it is 
gratifying to me to observe the confidence which is always placed in the 
correctness of the time as indicated by that gun. 
True, there are occasions when it misses fire altogether, and it has 
been known to go off of itself, or fire to a false signal preceding the signal 
which is intended to discharge it. Such cases are, however, rare, and the 
resulting uncertainty when they occur is usually so great as to deceive 
nobody. 
This gun, as most of you know, is fired by an automatic electric signal 
transmitted from the normal mean time clock at the Eoyal Observatory, 
the same signal being also transmitted to the Docks, where a ball is 
dropped, to Simonstown, to Port Elizabeth, and to East London. 
An automatic return signal from the Dock Time Ball, as well as 
observations both of the flash and report of the gun, and monthly returns 
from outlying stations, indicating the time shown by reliable chrono- 
meters at which the signals are received, combine to assure us at the 
Observatory that, except in cases of accident, the signals actually received 
at these distant stations correspond in time with those sent from the mean 
time clock. 
As this daily signal governs the whole of the time service — railway 
time, telegraph time, &c. — it is incumbent on the Observatory staff to 
ensure that the indications of the clock should not be merely regularly 
transmitted — a responsibility which, after all, attaches more closely to the 
telegraph department which provides for their transmission — but that they 
should be correct indications of the true statutory time. It would be 
quite safe to say that, with the precautions adopted for this purpose, the 
error in these signals never attains the limit of half a second, and but 
rarely reaches one-fifth part of a second. 
Now, no existing clock could be relied on, however carefully mounted 
and adjusted, to secure this accuracy for any continued length of time, and 
it is therefore necessary to have recourse to daily astronomical observa- 
tions to ascertain the errors of our clocks, which may then either be 
corrected or allowed for, as may be most convenient. In fact, the only 
timekeeper on which we can sufficiently rely is the earth itself in its 
(diurnal rotation, and the most direct method of accurately reading this 
