294 
Proceedings of the Royal Society of Edinburgh. [Sess. 
After iron has once been magnetised and demagnetised by decreasing 
reversals all of these assumptions no longer hold, however perfectly the 
process may have been carried out. The process of demagnetisation, 
however, is such that for reasons of symmetry any unit area on the surface 
of the hypothetical sphere contains an equal number of N. and S. poles. 
Further, the distribution and stability of the molecules must be symmetrical 
in reference to the directional demagnetising force, i.e. zonal. For different 
parallels of latitude either the distribution or the stability or both must 
vary. 
Consider (fig. 1), the stereographic projection of a hemisphere on the 
plane of the paper containing the demagnetising force ab and the sub- 
sequent magnetising field H H. The dotted curves are parallels of latitude 
in reference to ab, which may now be supposed to have demagnetised 
the iron by decreasing reversals. The little circles represent the equal 
distribution and stability of the molecules along these parallels. The 
numerical values to be given to these circles for different parallels need 
not be considered at present. Meanwhile let each little circle represent 
one molecule of N. and one molecule of S. polarity of known latitude 
and longitude in reference to ab. When referred to the polar axis H H 
the new latitude (or the complementary polar distance) and longitude 
of each molecule can be calculated, or read off if Mr Blaikie’s device for 
solving spherical triangles be used. A modification * of this apparatus 
facilitated the calculation involved. Let it be assumed that on the 
application of the positive field + H, each molecule leaves its known position 
and takes up a final position which, on the average, lies on the same 
meridian. Each N. and S. molecule experiences, after its rotation towards 
its final position under any given force + H, a transverse vertical change 
of orientation which will be zero for those molecules rotating along the 
zero meridian (horizontal), increasing, however, according to the sine law 
as the great circles through which the molecules rotate approximate to 
90° on either side. If the rotation of the N. molecules towards (up) and 
from (down) the equator at right angles to H H, in the upper half of the 
hemisphere be regarded as positive and negative respectively, in the 
* A circular disc of cardboard, on which was drawn the stereographic projection of a 
hemisphere, could rotate from beneath on the surface of a wooden board. Over this 
arrangement a piece of transparent paper was securely stretched. Obviously the new 
co-ordinates of any series of known points marked on the paper could, on the rotation of 
the disc through any required angle, be immediately read off. Mr Blaikie’s apparatus 
has been exhibited at the Koyal Society of Edinburgh and other societies, but no account of 
his ingenious device appears to have been published. Dr Dyson, Astronomer Royal, kindly 
showed me Mr Blaikie’s device at the Royal Observatory, Edinburgh, which suggested, 
the above modification of the more complete apparatus. 
