298 Proceedings of the Royal Society of Edinburgh. [Sess. 
parallel, and the magnetising field positive. The seven double columns 
correspond to the abscissae of fig. 2 from zero to T25 and represent values 
of the magnetising field. The higher values between 1*25 and 4 are 
omitted on account of the space that would be required. The polar 
distance and longitude of the equally spaced molecules on the zone 
selected (fig. 1) are determined when referred to the polar field +H. The 
longitudes are entered in the vertical column on the left, the polar 
distances in the zero column above referred to. Reference to fig. 2 
determines for increasing ordinate values the polar distances of the 
molecules, which are entered in their respective columns *2, *4, ‘6, etc. 
The sines of these angles into the sines of the longitude occupy the adjacent 
columns. The summation of these, each diminished by the summation of 
the first column (original position of the molecules), gives the change of 
the transverse components of the molecular moments for the various 
values of the horizontal ordinates (positive field) increasing from zero. 
The lowest horizontal column gives the summations of the upper and under 
halves of the zone. A zone, however, may occur in one half of the hemi- 
sphere only. 
Apparatus. 
The thin iron cylinders previously used, formed from sheet transformer 
iron, were in these experiments replaced by a length of drawn steel tube. 
The steel tube, inserted into a solenoid of covered copper strip, surrounded 
a bundle of copper wires along its axis. By means of these conductors 
the tube could thus be subjected to a circular, a longitudinal, or, when 
the central wires and solenoid are joined in series, to a spiral magnetis- 
ing force. This last constitutes the demagnetising force, and may be 
inclined at any angle 0 to that due to the central wires which con- 
stitutes the subsequent demagnetising field H c . An exploring coil of 
many turns of covered copper wire surrounding the central portion 
of the steel tube measures the longitudinal, i.e. the transverse induction 
change. 
The circular induction B c is wholly completed in the steel, while the 
longitudinal induction completes its magnetic circuit in part through the 
steel and in part through a massive iron yoke. The reluctance of the latter 
is, owing to its large sectional area relative to that of the steel tube, negligible 
so long as its permeability remains high. When the longitudinal induction 
is great, and consequently the permeability of the yoke reduced, this induc- 
tion in the tube may fall below its calculated value. A correction may thus 
