12 THE ANTARCTIC MANUAL. 
Il. 
PENDULUM OBSERVATIONS. 
By R. T. Guazesrook, F.R.S. 
THE apparatus is arranged in order to measure accurately the times 
of vibration of three pendulums. The period of each pendulum 
is about half a second, and these periods are determined by com- 
parison with a clock or chronometer. 
The main observing station will be adjacent to the winter quarters, 
and here the clock can be used. If opportunity occurs for making 
observations when at a distance from the base, the chronometer is 
available. 
Observations at the Base—Since the mean rate of a clock over 
24 hours may often be appreciably different from its rate for a short 
interval between the times for which the mean rate is determined, it 
is desirable that, wherever possible, (1) the pendulum observations 
should extend over the 24 hours; and (2) that astronomical obser- 
vations to determine the rate of the clock should be made at the 
beginning and end of that period. 
To secure (1) each pendulum should be swung for 8 hours, while 
to secure (2) the observations should be begun at a time at which 
star transits for the determination of the time are possible. 
A complete set of observations will occupy more than the 
24 hours, but much valuable information can be obtained without 
taking a complete set on every occasion. Moreover, it must-be noted 
that observations are not continuous during the 24 hours. When the 
apparatus is properly adjusted readings at intervals of 8 hours suffice, 
though at first a somewhat shorter time than this should be allowed 
to elapse between the observations. 
The method of coincidences is used to determine the periods. 
Light from a lamp is reflected through a narrow horizontal slit 
towards the pendulum-stand. A mirror attached to the upper part of 
the pendulum reflects the ray back, and it is viewed through a tele- 
scope which is adjusted so that when the pendulum is at rest the 
