DR. J. HOPKINSON ON THE MAGNETISATION OF IRON. 
459 
the soft iron bar to the body of the block. A small correction is required, important 
in the case of bodies but slightly magnetic, for the fact that the area of the exploring 
coil is greater than the area of the bars tested. Thus the induction measured by the 
exploring coil is not only that in the sample, but something also in the air around the 
sample. The amount of this was tested by substituting for a sample of iron or steel a 
bar of copper, and afterwards a rod of wood, and it was found in both cases that the 
induction 03 was 370 when the force <£) was 230. The correction is in all cases small, 
but it has been applied in the column giving the maximum induction, as it materially 
affects the result when the sample contains much manganese, and is consequently very 
little magnetic. 
The resistances were determined by the aid of a differential galvanometer, the 
connections being made as shown in the accompanying diagram, fig. 3. The additional 
resistance It was adjusted until the balance was obtained. The resistances actually 
measured are, some of them, as low as goVo °f an ohm, they must not therefore be 
regarded as so accurate as determinations made upon samples of a more favourable 
form; they, however, do show the remarkable effect of several impurities in iron, 
though it is possible that some of the results may be in error nearly 1 per cent. 
Results obtained. 
In all, thirty-five distinct samples were tested, of twenty compositions. The first 
three were supplied to me by Messrs. Mather and Platt, and of these I have no 
analyses. All the rest were analysed for me in the laboratory of Sir Joseph 
Whitworth and Co., and the samples of material were actually prepared there, 
excepting, Hadfield’s steel, No. X.; Bessemer iron made by the basic process for 
telegraph wire, No. IV., from the North-Eastern Steel Company; and two Tungsten 
steels, Nos. XXX. and XXXI., which are in general use for permanent magnets. 
I would express here my great indebtedness to Mr. Gledhill, one of the managing 
directors of Sir J. Whitworth and Co., for preparing for me the samples I desired, 
and having them analysed. The fact is, indeed, that any value this paper may 
possess really lies in the variety of samples tried and in the accompanying chemical 
analysis, both due to Mr. Gledhill. Samples Nos. I.-X. and XXXII.-XXXV. were 
tested with a pair of bars, the rest with a single bar of the sample used, in com¬ 
bination with a bar of wrought iron. The particulars of the several samples are 
most conveniently given in a table which follows, and to which I shall presently refer. 
With many samples observations were made sufficient to plot the ascending and 
descending curves which express induction in terms of magnetising force, but as these 
can make no pretence, for reasons already stated, to such accuracy as would warrant 
their use in testing a theory as to the form of curves of magnetisation, a few only are 
given as examples, and in other cases results are given in the table sufficient to define 
