398 Quartz-Thread Vertical Force Magnetograph. 
optically plane. The platinum stirrups for the magnets 
having been fitted, the small silica upright for the inertia-bob 
is fused to one of the tags. The surface of the mirror having 
been platinized, the magnets are fitted in.place, a small 
quantity of fused shellac being used to prevent the stirrups 
shipping. 
To prepare the fibres, a rod of fused silica having a dia- 
meter of about 2 mms. is taken and heated in a small oxy- 
hydrogen flame, and then rapidly drawn out. Two stops 
must be arranged so as to limit the distance the hands can be 
separated, and hence the length of the fibre produced. With 
a little practice it is possible to obtain a fibre of a suitable 
length and diameter (0:008 to 0°010 cm.). The two portions 
of the original rod must be cut off about a centimetre from 
the end of the fibre. One of these ends is fused to one of the 
tags attached to the mirror. The other end is in one ease 
fused to the quartz spring, and in the other is soldered by the 
process described by Threlfall* to a small brass rod which 
fits into the clamp attached to the screw H (fig. 1). 
To adjust the centre of gravity, the movable system is 
mounted between two uprights on a wooden board, while a 
small U-shaped piece of brass, with slots to take the tags 
attached to the mirror, supports the mirror and magnets. 
The rods of quartz between the fibre and the mirror are then 
softened by heating in a very small oxyhydrogen flame, and 
bent so that the centre of gravity lies a little below the line of 
support. During this operation the uprights of the U protect 
the magnets and mirror from the heat of the flame. The 
operation is one requiring some delicacy of touch, but is really 
not as difficult as one would suppose. 
With reference to the adjustment for temperature compen- 
sation, the side on which the weight will have to be placed 
depends on which side of the suspension-line the centre of 
gravity is situated, since it is almost impossible to get it 
exactly vertically below this line when adjusting the system. 
By, however, running the instrument in a situation where the 
changes of temperature are considerable, it will soon be seen 
whether it is over- or under-compensated. By making two 
sets’ of observations with the weight in two widely different 
positions, and noting the amount of twist put into the fibre, the 
position for compensation can be calculated. In this con- 
nexion the easiest way is to calculate the correct amount of 
twist for the fibre, put in this amount of twist, and then adjust 
the weight till the magnet is horizontal. 
* ‘Laboratory Arts,’ p. 226 (Macmillan & Co.). 
