444 PKOFESSOR W. WATSON OX A DETERMIXATIOX OF THE VALUE OF 
The value of F changes slightly with temperature, and assuming the coefficient of 
expansion of the coils to be 0'0000187, a table of the values of F for each degi-ee of 
the range of ternpei’ature met with duriug the observations was drawn ujd. 
A consideration of the figures given in the tables of measurements will show that, 
as might be expected, the greatest variations occur in the measurements of the 
external radii of the coils. Taking into account the fact that the internal radii can 
be determined with more accuracy, the mean radius ought not to differ from the 
truth by as much as 0'005 centim. The distance between the mean planes is 
probably known to within about the same amount. Thus the uncertainty ui F 
appears to be less than 5 parts in 30,000. This corresponds to an uncertainty in H 
of about 3 in the fifth place. 
The adjustments which have to be made before the coils are used are as follows :— 
1. The axis of the coil must be horizontal, and the axis about which it turns 
vertical. 
2. The axis of the coil must be perpendicular to the magnetic axis of the 
magnet. 
3. The centre of the coil must lie on the vertical axis of the magnetometer, and 
the magnet must be at the centre of the coil. 
The first of these adjustments was made by means of a striding level which rested 
on the flanges of the coils. The magnetometer was first levelled so that the reading 
of the striding level remained the same when the coils were rotated. When this 
was complete the axis about which the coils turned was vertical. To make the axis 
of the coils horizontal, the screws which bear on the deflection bar (D, fig. 1) were 
adjusted till the level reading remained the same when the striding level was 
reversed. 
In making adjustment 2 it was at first assumed that the magnetic axis of the 
needle was perpendicular to the mirror attached to the needle. If this be so, the 
axis of the coil will be perpendicular to the axis of the magnet when it is so adjusted 
that it is perpendicular to the line joining the axis of suspension of the magnet to 
the zero of the scale, S fig. 1. It was ascertained by experiment that this line was 
at right angles to the axis of the small tube which is used in the magnetometer for 
adjusting the height of the deflected magnet when this tube was laid in the V’s 
used for supporting the magnet in the deflection experiment. Thus it was sufficient 
to make the axis- of the coil coincide with the axis of this sighting tube. The 
position of the coil was adjusted by means of the screws which bear upon the 
magnetometer till the two crosses on the mica discs, which indicate the centres of 
the two coils, were on the axis of the sighting tube. Separate experiments were 
made in the case of the two magnets used in the experiments to measure the angle 
included between the mao-netic axis of the inannet and the normal to the mirror. 
o o 
In the case of the longer magnet this angle was 2', while in that of the shorter 
magnet it was 38'. As the correction which would have to be applied, owing to the 
