MAGNETIC SHIELDING IN HOLLOW IRON CYLINDERS. 653 



There being no reason to suppose that the molecular magnets topple over more in one 

 direction than another, it follows that the molecules will tend to point less towards 

 a zone making an angle of 45° with the magnetising force, more towards the magnetis- 

 ing or polar force H, and also more towards the equatorial regions. (See diagrams 1 

 and 2, fig. xix.). We may therefore conclude that on the whole, owing to the greater 

 number of groups on the equatorial side of the 45° latitude line, the molecular magnets 

 will tend to point in a plane or equatorial belt at right angles to the demagnetising 

 force. But the pull exercised by the demagnetising force H is greatest on those 

 molecules lying equatorially ; hence it follows that if the direction of the subsequent 

 magnetising force be rotated, the number of molecules lying equatorially is reduced and 

 the pull becomes less and less as the rotation approximates to 90°. Hence we have 

 obtained a possible explanation, in terms of molecular theory, why iron is not 

 isotropic to a magnetising force at right angles to that used in the demagnetising 

 process. 



§ 36. At this stage a further experiment suggested itself. The iron shields were 

 thoroughly demagnetised, first by decreasing reversals of H, and then of H c , and, second, 

 subjected to a maximum magnetising force H h which on being withdrawn left the iron 

 with residual magnetisation. This amounted to B^ = 760 C.G.S. units for shield A and 

 to B t = 800 C.G.S. units for shield B. The last four columns of Table XIII. supply the 

 data when the B-H curve due to H c increased by increments is superposed upon this 

 residual magnetisation at right angles to the magnetising force. In fig. xviii. these 

 measurements are plotted as dash line curves (3), and at the earlier stages of induction 

 the increase of the permeability is marked when compared with either of the two other 

 curves. We have thus for the two qualities of iron experimented with a sequence of 

 three curves, all obtained by increasing the magnetising force H c by increments from 

 zero. In each case, however, the iron had been left in a different molecular condition, 

 unsupported by any external magnetic field. If we consider the full line curves 

 obtained by the re-application of the same directional force as that used in the 

 immediately preceding demagnetising process, the usual induction curve under normal 

 conditions ; then iron demagnetised by a force at right angles to that subsequently 

 applied is less permeable, and iron left with residual polarity due to a force at right 

 angles to that subsequently applied is more ^permeable than iron magnetised under the 

 normal conditions. Fig. xix. shows graphically and without further explanation why 

 the pull on the molecular magnets, and consequently the permeabilit) 7- at low values of 

 H, should be in an ascending order for curves 1, 2 and 3 (fig. xviii.). 



§ 37. In conclusion, it must be borne in mind that a possible explanation of the 

 general nature of the effects is all that is attempted. Thus in discussing the molecular 

 condition of iron demagnetised by decreasing reversals, it has been assumed that H = K 

 cos 0. A different law for H would give a different angle for the maximum turning 

 moment. Perhaps a point worth noting is this, that unless it can be shown that the 

 turning moment is the same for all values of the angle 6 during the process of 



