20 Barrett, Brown & Hapriretp—On the Magnetic and Electric 
magnetic difference between the silicon-iron and aluminium-iron is seen at once 
by comparing Plates III. and IV. Up to this we have had no ring of aluminium- 
iron so that the above results can only be considered as comparative. If we take 
the figures obtained from the rings of Swedish charcoal-iron and silicon-iron, we 
see, from Tables VII. and XII., that, at a magnetising force of about 60 units, the 
maximum induction, coercive force, and hysteresis loss of silicon-iron (898 E) 
comes between S. C. I. and 1167 H.* 
It seemed desirable to make a more strictly comparative experiment of the 
relative values of these three substances at a maximum induction of about 9,000, 
which is the highest limit of the induction usually employed in transformers. 
The rods were therefore re-annealed simultaneously; and the cycles shown on 
fig. 8, on previous page, were taken by the magnetometric method. From these 
curves the following magnetic constants were deduced, and for the sake of 
comparison the results usually given for good soft iron are included :— 
TABLE XIII. 
Hysteresis Losses ror MAxrmum Inpucrion = 9,000 C. G. S. anp Frequency = 100 
or Rops or Sixrcon-Iron (898) anp Atuminium-Iron (1167) comparED wirtt 
ORDINARY AND PURE SOFT IRon (S. C. I.) ALL ANNEALED. 
Hysrerusis Losses. 
Watts. 
Hysteretic Microwatts | 
Mark. Density. Constant per ¢.c. Per cubic inch. | Per lb. 
Soft Iron 7:78 002 42,440 0-695 | 2°50 
Sb OL IL 7°84 °0011 | 23,340 0°382 1°35 
898 K. 7°66 00078 15,490 0:254 0:92 
1167 H. 756 00068 | 14,480 0:236 0:86 
The energy dissipated is given in watts per cubic inch of material, and also 
per lb. per 100 cycles; and it will be seen that at this lower induction the 
aluminium-iron is considerably better than our purest specimen of iron. It is 
certainly an extraordinary fact that, adding 21 per cent. of a non-magnetic 
element to iron, should confer upon the iron a higher magnetic quality. This 
is probably due to the increased mechanical softness given by alloying iron 
with aluminium. 
* The character of the cyclic-curve for 1167 His somewhat peculiar; it swells out at the bend more 
than the 8. C. I., and contracts more rapidly than the iron as the horizontal axis is approached. 
