PENDULUM OBSERVATIONS. 15 
the standard. This may prove to be rather low in the present case, 
but we will assume that it is the pressure aimed at. 
It is impossible, however, always to secure this pressure exactly, 
and thus a correction is required to reduce readings to the standard 
pressure. For this purpose the pressure p and the temperature ¢ are 
observed. To reduce the pressure y to what it would be at 0° we 
divide it by 1 + ‘00367 ¢, and obtain as the difference from the 
standard the value of 60 — SELES 
Now observations show that for a small range of pressure the 
correction to the period is proportional to the change of pressure: 
if we call &' the increase of period produced by a pressure-change of 
1 mm. the correction due to a change of 60 — ane will be 
Pp ) 
1+ -00367 tj * 
observations before the start. 
It must be remembered that while accuracy is secured by working 
at a low pressure this is not absolutely essential to success, though 
the time over which the observations can be extended will be seriously 
reduced if the pressure be high. At the same time should it be 
impossible for any reason to maintain the low pressure, experiments 
made under a greater pressure would, if carefully conducted, give 
important information. In all cases the pressure at the time of the 
observation must be carefully recorded. 
Are Correction.—The period depends on the amplitude of the 
surveys of the pendulum. The correction on this account is given 
if the amplitude of the first and last surveys can be found. 
To obtain these a vertical scale is attached to the front of the box 
which contains the shutter. By shifting the lamp used with the 
flash apparatus this scale can be illuminated, and its reflection on 
the pendulum mirror made visible. By reading the divisions of the 
scale which coincide with the cross wire at the extremities of the 
swing, the amplitude is given, if we know the distance between the 
scale and the mirror. In calculating the amplitude it must be 
remembered that the reflected ray is turned through twice the angle 
through which the mirror moves. 
In the instrument supplied the small divisions are about 3 mm. 
in length, and at a distance of 2 metres the angle subtended by a 
division is very nearly 5’. If the amplitude of the oscillation be 
one division the angular amplitude of the swing of the pendulum 
will be 2'°5. 
i 00 - The quantity #’ will be determined by 
