Properties of an Extensive Series of Alloys of Iron. 9 
the long slender reds, S. C. I., 898 E, and H, were turned down to 0°4 em. diameter 
and 102 cms. long. Unfortunately 898 H broke in the lathe, and hence in this 
third series of experiments it could not be compared with 8.C. I. After being 
turned, both rods were carefully annealed by heating to whiteness in a gas-flame and 
cooling slowly in an east-and-west position. Annealing was also carried out by 
packing the rods in an iron tube with lime, heating to whiteness in a combustion © 
furnace, and allowing the whole to cool slowly during the night: the magnetic 
measurements taken after this method of annealing did not, however, show much 
was gained in permeability. Plate II. shows the cyclic curves of the 24 per cent. 
silicon-iron alloy, 898 E, compared with that of our best iron S.C.I., both rods 
being 255 diameters in length; the correction for the reaction of the ends, 
estimated from Professor Hwing’s experiments, is shown by the slope of the vertical 
axis: that is to say the value of the magnetising force H, required to produce 
any given induction B, must be estimated from the sloping line, and not from 
the vertical. The magnetometer employed was a delicate reflecting one, placed 
70 cms. from the pole of the test rod, the needle being suspended by a fibre 
of unspun silk. It was found that the pivoted magnetometer employed in the 
earlier part of these researches (see Trans. R.D.S., vol. vii., p. 67) had become 
sluggish from wear, and a series of cyclic curves of these rods obtained by its 
means had to be rejected, as they proved to be untrustworthy.* 
A fourth set of experiments for the purpose of testing this silicon-iron was 
made with rings of the alloy, using the ballistic method, undoubtedly the best 
but the most tedious. Castings were made of 8. C.I. and 898 E; these were 
forged, and then annealed. The blocks were turned into rings in the Laboratory 
of the College of Science, and after turning raised to a bright red heat in sand, 
and allowed to cool slowly in the sand; another specimen of 898 E, after being 
turned into a ring, was sent to Sheffield for annealing. The dimensions of the 
rings were carefully determined and are given in Table IV., the cross-section 
TABLE, TY. 
Mean Diameter. Cross sectional area. Density. 
Sb Ch Ile 9:022 cms.. 0:7299 sq. cms. 7842 
898 EK > GORY on 0:7649 =, 7°665 
* That is, the cyclic curves when plotted were all found to be flattened in the part where the induction 
changes rapidly with change of field; the relative value of the curves, however, was not much affected. 
Our experience shows that it is not safe to trust even the most delicate pivoted magnetometer in 
magnetic measurements where successive small changes of force require to be measured. 
TRANS. ROY, DUB. SOC., N.S., VOL. VIII., PART I. C 
