RESEARCHES IN MAGNETISM. 
575 
rotation of the magnetic molecules, we should be obliged to assume either that this 
friction was vanishingly small for some of the molecules, or that a certain range of 
motion was possible, to all or some of them, before the frictional resistance began to 
be felt. We have seen already that one of these assumptions would be required to 
account for the indefinitely small retentiveness which hard iron and steel seem to 
possess when exceedingly feebly magnetised. 
But though the observed values of k extend to very low values of 2>, between these 
and the zero of magnetisation there is room for a complete change in the form of the 
curve, and it would be rash to infer anything respecting the initial value of k from the 
way in which the curve is trending at the finite magnetisations for which the lowest 
observed values of k have been determined. The curves, indeed, give some slight 
evidence that a downward bend occurs near the origin, and, in any case, they afford 
no positive proof that k is not initially zero, or even negative. 
§ 62. If we could suppose that the frictional sticking of the molecules was at first 
complete, so that none of them budged from their primitive positions wdien a very low 
value of Q was applied, then we should expect to find the initial value of k not zero, 
but a negative quantity. For in that case the cause producing paramagnetism 
(namely, the rotation of the molecules) would be entirely inoperative, and the applica¬ 
tion of «§ might be expected to induce Amperian currents in the stationary molecules 
in the same manner as in other substances, with the result of making the magnetisa¬ 
tion opposite in direction to the magnetising force, until the latter became large 
enough to overcome friction and begin turning the molecules. Thus, if none of the 
molecules could begin to turn until <§ reached a certain finite value, we should expect 
to find iron diamagnetic with respect to lower values of <§. 
[Note added April 9, 1886.—It is possible that substances which are diamagnetic 
in fields of ordinary intensity are so, not because the molecular currents are absent or 
weak, but because the molecules are held so fast that their alignment (which would 
give paramagnetisation) is produced with difficulty. So long as they do not turn, or 
turn very slightly, diamagnetic induction is the resultant effect of the applied field. 
If the resistance to the turning of the molecules resembles that found in iron, we 
should expect the diamagnetic induction which occurs when the field is moderate to 
give way to paramagnetic induction as the field is increased ; also that a residual 
paramagnetic polarity would be left after a strong field has been applied and removed, 
4 e 2 
