repetition of M. Arago's experiments on magnetism, &c, 495 
side, and be again renewed on the side in advance, but 
at first not in its full intensity ; so that it is not merely the 
diminution of surface arising from the abstraction of a part of 
the metal, but a much more considerable defalcation of mag- 
netic force which takes place on either side of the slit, that 
operates. Now this operation is always to weaken the drag 
between the magnet and the disc, and no reason, a priori, can 
be assigned why this effect should not take place to any 
extent. 
39 The validity of this reasoning is shown by taking the 
extreme case in which the substance acted on is in the state 
of powder. Each particle of this becomes necessarily a feeble 
magnet, and its north and south poles, being at the same 
distance (almost precisely) from the pole of the magnet, 
counteract each other’s action. The extreme feebleness of 
their magnetism prevents the particles from affecting each 
other by induction across the intervals which separate them ; 
so that each acts as an individual, and destroys in great 
measure its own effect. The moment’ however a metallic. 
1 ' 
i. e. a magnetic contact is established between them, their 
mutual induction acts, and the result is a general develope- 
ment of one polarity in the region adjacent to the magnet ; 
and of the other, feebler and more diffused, in the parts of 
the mass remote from it. This is probably the rationale of 
the restoration of virtue which takes place when a cut disc is 
soldered up. And it is not difficult to conceive that a weak 
magnetism may be thus very faithfully transmitted through 
substances, such as bismuth and lead, whose direct action is 
very small, because, as we have seen, the intensity of their 
direct action depends, for one of its causes, on the retentive 
