240 M. Werner Siemens on the 
from experiments to be described later on, the total magnetism 
in an interrupted ring must be less, corresponding to the great 
increase in resistance to the magnetic disturbance caused by 
the air-filled space of the gap ; and the disturbance-lines, or 
lines of force which bind together the parts of the ring, must 
occupy the whole surrounding space in very different inten- 
sities, and produce in it the phenomena of magnetic attraction 
and distribution, or those of the so-called free magnetism. 
Hence the Amperian theory must be extended by supposing 
that not only magnetic substances, but all bodies as well 
as empty space are filled with circular currents of very small 
dimensions, and that magnetic substances differ only from 
nonmagnetic substances inasmuch as the number of circular 
currents present in the unit volume is much greater in the 
first case than in the second. 
All magnetic phenomena may then be referred to the pro- 
perty of the electric current of exerting a directive force upon 
the molecular solenoids which fill all space, but which are 
present in greater numbers in the so-called magnetic bodies, 
which place their axes at right angles to its direction and 
tend to bring them into closed concentric attraction-circles. 
The magnitude of this rotation of the axes depends, on the 
one hand, on the magnitude of the directive or magnetizing 
force, and, on the other hand, on the number of the molecular 
circuits preexisting in the unit-volume, for which condition 
the term ‘‘ magnetic conductivity ’’ may be employed, or that 
of “ magnetic resistance’ for its reciprocal value. 
To test the admissibility of this view, the case of an iron 
ring, or closed ring of iron tube, uniformly wound with msu- 
lated wire, which case has been repeatedly examined experi- 
mentally and theoretically, seemed to me especially suitable, 
since, according to G. Kirchhoff’s investigations, such a ring 
produces no magnetic effect at a distance. 
For the magnetism of an iron tube having a thickness of 
wall s, traversed by an axial current, I have previously* 
obtained the value 
M=47lsi 
from Ampére’s formula—where s is the thickness of metal, 
Lihe length of the tube, and 7 the strength of the current—and 
have shown its correctness by means of experiments. 
If we surround an iron ring of sectional area g and mean 
radius p with a closely wound spiral, then, in accordance with 
the above considerations, the magnetizing force is proportional 
* Sitzungsber. d. Berl. Akad. 1881, p. 701. 
