302 
Proceedings of tlie Koyal Society of Edinburgh. [Sess. 
previously demagnetised by the spiral force making a calculated angle 
of 0 = 18 o, 4 with H c . If demagnetisation takes place at a calculated 
value of 0 = 11 6 (complementary angle), the broken line curves result. 
The orientation of the spiral demagnetising field is indicated by continuous 
and broken line arrows for 0= 18°*4 and 7l°'6 respectively. If the tangents 
of those angles, given by the ratios between the calculated values of the 
longitudinal and circular fields, be positive, the spiral demagnetising force 
may be considered right-handed, and if negative left-handed. 
If the demagnetising force be right-handed, the transverse induction 
change occurs in the first or third quadrants according as the subsequent 
field is positive or negative. On the other hand, if the demagnetising 
force be left-handed, the transverse change occurs in the fourth or second 
quadrant as the subsequent field is positive or negative. 
In this respect, therefore, the results previously arrived at are con- 
firmed. The curves, however, are more symmetrical in the four quadrants 
than those previously obtained without the massive yoke. 
The curves plotted in fig. 4 are typical of a series of experiments 
for other calculated angles of the demagnetising force, giving equally 
symmetrical results. Fig. 5 gives the complete series, but the average of 
the readings obtained in the four quadrants are plotted in the first 
quadrant, the demagnetising spiral force being taken as right-handed 
and the subsequent field (H c ) positive. The highest curve, so far as 
observed, is that obtained when the demagnetising angle is 0 = 45°. When 
rotated in either direction the curves take lower values, and finally vanish 
as 0 approximates to 0° or 90°. 
It will be noted that the curves obtained for pairs of the calculated 
complementary angles do not coincide. The angle at which demagnetisa- 
tion takes place is determined by the actual ratios of the longitudinal 
and circular inductions in the shell of the steel tube. It was found 
experimentally that, when the longitudinal component of the combined 
inductions remained small, and consequently the magnetic reluctance of 
the iron yoke negligible, the angle at which demagnetisation actually took 
place coincided with that calculated from the field ratios so nearly that no 
correction is required when (9 = 9°*4, 18 0, 4, and 26°*5. 
On the other hand, it was found that when the longitudinal component 
was large, and consequently the magnetic reluctance increased at high values 
of the combine^ fields, the angle at which demagnetisation took place did 
not remain constant during the demagnetisation process. Fig. 6 shows the 
variation which occurs as the spiral alternating induction decreases from a 
maximum of B = 15,000, as determined by the ratios of the longitudinal 
