38 Scientific Proceedings, Royal Dublin Society. 
should synchronize with the central or standard clock, and show as 
nearly as possible the true time. 
Under these circumstances I made the following recommenda- 
tions :— 
The installation naturally divided itself under three heads— 
(A)—A. standard clock, of best procurable make, placed in such a 
position as would ensure its protection as far as possible 
from climatic or other changes, and furnished with elec- 
trical connexions suitable for the control of — 
(B)—A strong piece of uniform motion clockwork automatically 
controlled at stated intervals (say, every 15 seconds) 
from A, and further checked and controlled once in 
every 24 hours by the Greenwich signal received through 
the Post Office: this piece of mechanism to be enclosed 
in a glass case exposed in some prominent position (say, 
in the entrance hall), supplied with dials and hands, and 
arranged also so as to serve as a distributing clock for 
driving. 
(C)—A. series of, say, 200 dials distributed throughout the 
building, each being furnished with minute and hour 
hands, and the necessary dial work, actuated once a 
minute only from the distributing clock B in the entrance 
hall. 
The distributing clock B in the hall, being automatically con- 
trolled once in every 15 seconds from the standard clock, may be 
depended upon to keep as accurate time as the standard clock 
itself, hour by hour; while, to prevent any accumulation of small 
errors which might, in the course of some days, reach an incon- 
venient amount, there is the further check by Greenwich signal at 
10 a.m. or 1 p.m. every day. Any error that exists on receipt of this. 
signal is not only wiped out at once automatically, but the amount 
of this error is registered on a paper tape for after-reference, if 
desired. 
The maximum error that can exist, therefore, cannot be greater 
than the rate + or — of the standard clock for 24 hours; while, after 
the receipt of the Greenwich signal, the clock B may be relied on 
to within a fraction of a second. 
