18 Barrett, Brown, & Hapristp—On the Magnetic and Electric 
the middle part of the rod to be tested; this coil consisted of 400 turns of fine silk- 
covered copper wire in 4 layers. The rod and test coil were now put into the long 
magnetising helix (about 20 ems. longer than the rod) having a magnetising 
force = 30:°2 C.G.S. unitsper ampére. The ends of the secondary coil were joined 
to the ballistic galvanometer through an Earth inductor in the usual way, and a 
series of steps taken in the same manner as in a ring test. The first parts of the 
B-H curves for the aluminium-iron alloy and the Swedish charcoal-iron are shown 
in fig. 7, and the values of the induction and permeability in different fields, 
deduced from fig. 7, are given in the next table. 
TABLE XI, 
Rops 260 Diameters In LENGTH TESTED BALLISTICALLY. 
Swedish Charcoal Iron. Aluminium-Iron Alloy. 
H B ip H. | B 
‘ =; af : 
2 6,000 3,000 2 | 10,850 5,425 
4 11,600 2,900 4 | 18,600 3,400 
6 12,900 2,150 6 | 14,300 2,360 
| 8 13,550 1,690 8 | 14,650 1,830 
| 10 14,000 1,400 10 | 14,950 1,495 
If this table be compared with Table IX. the results will be seen to be fairly 
concordant; but no correction has been made for the demagnetising reaction of 
the ends in the ballistic method ; this correction would, of course, raise the above 
values of the induction and permeability.* 
Before the ballistic experiments were made, the values of the hysteresis loss in 
the case of the two rods was found by a number of complete cycles determined 
magnetometrically with the following results. 
The aluminium-iron rod of length = 260 diameters was put through a complete 
magnetic cycle up to a field of 60 units, and compared with a similar cycle obtained 
from the rod of Swedish charcoal-iron of the same dimensions, both having been 
subjected to identically the same heat treatment. The two cyclic curves are 
shown on Plate IV.; the sloping vertical axis indicating the correction for the 
demagnetising reaction of the ends. The aluminium-iron, 1167 H, shaded on the 
plate will be seen to have not only a higher permeability than 8S. C. L, but a greater 
* Tf the same correction be made in both methods, the figures for the ballistic method appear to be 
somewhat high ; indicating either that a lower correcting factor is necessary, or that the inductance per 
scale division of the galvanometer has been over-estimated in this method. 
