GREENWICH TI*IE AND ITS TELEGRAPHIC DISTRIBUTION. 361 
coil the properties of a magnet, and reversion of the direction 
of the current reverses the direction of magnetism. If, there- 
fore, a current is passed through the coil such as will cause 
attraction between the adjacent ends of the swinging magnet 
and fixed coil, the pendulum will be accelerated. An opposite 
current, producing repulsion, will retard the pendulum. Either 
current must remain in action (as at present arranged) ten 
minutes to change the clock by one second of time. An error 
of a small fraction of a second can thus be easily and certainly 
corrected. 
The reader will remember that we brought up the descrip- 
tion of the manner of determining time so far as to show how, 
by star observations, the error of the sidereal clock is obtained. 
Now the sidereal clock works, by galvanic means, a small 
clock placed on the desk of the Superintendent of the Time 
Department, near to the small mean-time clock before spoken 
of. On this desk, then, we have two small clocks — one repre- 
senting the sidereal clock, the other the normal mean-time 
clock, both of which are in other parts of the building. And 
between the two clocks is placed a commutator, by which a 
direct or reverse galvanic current can be passed through the 
coil in the normal clock for acceleration or retardation of its 
pendulum. To ascertain the error of the mean-time clock we 
take, at any instant at pleasure, by the sidereal clock, the read- 
ing of the mean-time clock. Knowing the error of the sidereal 
clock (we do not correct this error, it being more convenient 
for astronomical purposes to allow it to accumulate), the true 
sidereal time of comparison is known. The corresponding 
mean solar time is then easily calculated. The difference be- 
tween this calculated time and that read from the mean-time 
clock is its error, and the error also of the normal clock and 
all clocks in connection with it. The necessary correction is 
then made by turning the commutator handle to accelerate or 
retard the normal clock pendulum for as long a period as may 
be necessary : the rate at which the alteration is made is men- 
tioned in the preceding paragraph. The normal clock is thus 
corrected several times daily, always immediately before 10 h. 
a.m. and 1 h. p.m., for reasons that will be afterwards apparent. 
The correction required is usually some fraction of a second 
only. 
We have now to speak of a most important duty which the 
normal clock performs. This is the transmission of galvanic 
time-signals from the Observatory at the exact instant of each 
hour. For this there passes through the clock a galvanic 
circuit ; but in the course of the wire there are two breaks : 
one is united at the instant the seconds hand marks 60, the 
other from shortly before to shortly after the minute hand 
