568 
Proceedinys of the Royal Society 
Hence we find that the effect of any small excentricity will be a 
minimum with practically any value of h between 0 and 3, provided 
the coils are wound so that the ratio of a : ; C 2 is 10:2:10. The 
frame actually used has the following dimensions: — cm., 
5=1.8 cm., Cj = 1 . 8 cm., C 2 = 9 cm. 
[Independently of the above calculation, the deviating effect of a 
magnetic strip was determined experimentally. Upon a rectangular 
plate of wood a coil of wire was wound so as practically to represent 
an electromagnetic strip, with the following dimensions : — 
Internal width of the strip 
Internal height „ ,, 
External width ,, ,, 
External height „ „ 
Thickness „ ,, 
Humber of turns of wire 
= 0'37 cm. 
= 9-1 „ 
= 0-7 „ 
J. Vy X jy 
= 0*34 „ 
= 49 
A strong uniform current was supplied by four large gravity cells. 
The strip stood upright on two round legs in a V groove which 
was placed along the magnetic meridian, and 30 cm. east of the 
magnetometer, which hung in a horizontal plane through the bottom 
of the strip. To increase the sensibility, the directive field was 
considerably weakened without change of direction by using a 
permanent magnet. The apparatus is shown in fig. 9. 
Calling 6 the deflection of the magnetometer. 
tan 0 — 
.do) 
dx 
H + i 
jdoi 
dy 
with the usual notation. 
,doi . 
dy 
But since 
is small compared with H, we 
may take the deviating force as 
nearly proportional to tan 6. Fig. 
11 gives the curve plotted from this 
experiment. It will be seen that 
the fall after passing the maximum is rather rapid compared with 
that in curve III in fig. 8, which, allowing for change of scale, 
