496 Messrs. W. H. Preece and H. R. Kempe. [June 10, 



all it is, at all events, very much higher than that which was assigned 

 to it by Rowland.* 



There may, perhaps, be some doubt whether the expression 

 used above is exactly applicable to the case of my divided ring, 

 and small errors may possibly be introduced by the fact that the 

 contact between the opposite faces was not quite perfect throughout.! 

 But apart from minute accuracy of detail, the general character of 

 the results is entirely free from doubt, and would be quite unaffected 

 by a very large margin of uncertainty in the expression. They show 

 that the generally accepted ideas with regard to several important 

 points need modification. 



Thus it is not true that the lifting power of an electromagnet 

 reaches a practical limit under a comparatively small magnetising 

 force, and that even if excited by an infinite current it could not 

 support a weight of 200 lbs. per square inch of surface. 



It is not true that the magnetisation of iron becomes sensibly con- 

 stant when the magnetic force exceeds a certain moderate value. 



And it is not true that the maximum of magnetic induction, if it 

 exists at all, is represented by anything like so small a value as 

 18,000 units. 



In conclusion, I have to express my great obligation to Lord 

 Rayleigh for much valuable assistance and advice in the preparation 

 of this paper. 



VII. "On a New Scale for Tangent Galvanometers." By 

 W. H. Preece, F.R.S., and H. R. Kempe. Received May 6, 

 1886. 



Tangent galvanometers are much used for the exact quantitative 

 measurement of currents of considerable strength, such as are 

 measured in amperes, but they are not so generally used for the 

 measurement of smaller currents in milliamperes. This arises from 

 a notion that they are not sensitive enough; although the most 

 sensitive instrument in practice — Sir William Thomson's mirror 

 galvanometer — is really a tangent instrument. The ordinary forms 

 — Joule's, Gaugain's, or Helmholtz's — are not very sensitive. Their 



* It is hardly necessary to point out that if there is any limit to |I5, the suscepti- 

 bility, k, must become negative when the magnetic force exceeds a certain value. 

 Maxwell appears to have considered this not impossible. See " Electricity," vol. 2, 

 § 844. 



f The close agreement of the values of k in Table II with those obtained in the 

 experiments of Stolotow (so far as the latter go) affords strong evidence of the 

 accuracy of the method (see "Enc. Brit.," 9th edit., vol. 15 (1883), p. 255). 



