PROFESSOR TYNDALL ON THE DIAMAGNETIC FORCE, ETC. 
13 
in virtue of its structure, to set from pole to pole, it is opposed by the tendency of 
the longest dimension to take up the same position. Between the points the influ- 
ence of length is paramount, above and below the points the influence of structure 
prevails. 
Hence, the rotation of magnetic bodies, on being raised or lowered, is always from the 
axial to the equatorial position. 
The error of the explanation which referred many of the above actions to the pre- 
sence of two conflicting forces, one of which diminished with the distance in a quicker 
ratio than the other, lies in the supposition, that the assuming of the axial position 
proved a body to be magnetic, while the assuming of the equatorial position proved 
a body to be diamagnetic. This assumption was perfectly natural in the early stages 
of diamagnetic research, when the modification of magnetic force by structure was 
unknown. Experience however proves that the total mass of a magnetic body con- 
tinues to be attracted after it has assumed the equatorial position, while the total 
mass of a diamagnetic body continues to be repelled after it has taken up the axial 
one. 
III. ON THE DISTRIBUTION OF THE MAGNETIC FORCE BETWEEN TWO FLAT POLES. 
In experiments where a uniform distribution of the magnetic force is desirable, flat 
poles, or magnetized surfaces, have been recommended. It has long been known 
that the force proceeds with great energy from the edges of such poles : the increase 
of force from the centre to the edge has been made the subject of a special investiga- 
tion by M. VON Kolke*. The central portion of the magnetic field, or space between 
two such magnetized surfaces, has hitherto been regarded as almost perfectly uni- 
form, and indeed for all ordinary experiments the uniformity is sufficient. But, when 
we examine the field carefully, we find that the uniformity is not perfect. Substi- 
tuting, for the sake of convenience, the edge of a pole for a point, I studied the phe- 
nomena of rotation described in the last section, in a great number of instances, by 
comparing the deportment of an elongated body, suspended in the centre of the space 
between two flat poles, with its deportment when suspended betw'een the top or the 
bottom edges. Having found that the fibre of wood, in masses where form had no 
influence, always set equatorial, I proposed to set this tendency to contend with an 
elongation of the mass in a direction at right angles to the fibre. For this purpose 
thirty-one little wooden bars were carefully prepared and examined, the length of 
each bar being about twice its width, and the fibre coinciding with the latter dimen- 
sion. The bars were suspended from an extremely fine fibre of cocoon silk, and in 
the centre of the magnetic field each one of them set its length axial and conse- 
quently its fibre equatorial. Between the top and bottom edges, on the contrary, 
each piece set its longest dimension equatorial, and, consequently, the fibre axial. 
For some time I referred the axial setting of the mass, in the centre of the field, to 
* Poggendorff’s Annalen, vol. Ixxxi. p. 321. 
