1887.] Experiments on the magnetization of tron rods. 87 
Tests have been applied in very varied ways to eliminate 
sources of uncertain error. The cores were moved into different 
positions in the magnetizing spiral through distances considerably 
greater than accidental variation ; the induction coil was moved to 
different positions along the core—the alterations in the induction 
due to such changes led us to tie the induction coil firmly to the 
core; the cores were taken out and set afresh; the cut cores were 
pressed together by means of clamps. 
Experiments on bars of different diameters. 
The results recorded in Table I. were got for cores whose 
diameter was 12°7 mm., and of total length 10 cms. The experiments 
were repeated with cores of the same length but of smaller diameter, 
namely 33mm. The cut ends were carefully worked on oilstone, 
and the general results showed that but little difference was made 
by altering the dimensions of the core: or in other words, that 
there was the same surprising reduction for cores of small diameter 
as for cores of large diameter. When the separation of the worked 
ends was about 
‘4mm. the induction was reduced 26 per cent. 
8 mm. ” ) bP} 2 42 a) ”) 
33 mm. 2) ) o? ” Od ”? 2? 
Direction of the lines of force about the crevasse. 
Cores, cut as above described, were put into a long coil having 
two layers of wire, and fitting closely to the cores: and iron filings 
a an 
es | | 
Fic. 1. WoRKED ENDS TOUCHING. 
were sprinkled on paper laid horizontally in a plane through the 
a of the cores. Figures 1 and 2 were traced through the iron 
lings. 
It will be seen that there are fairly definite poles on either side 
