10 



PROFESSOR CARGTLL G. KNOTT ON 



it is impossible to say. It is more than probable, however, that it is largely referable to 

 the mode in which the magnetisation is applied. 



There is a second particular that appears to be an essential peculiarity ; and that is, 

 the way in which the permeability curve begins its ascent. It will be seen from the 

 table that the permeability has nearly the same value for fields below 0*05 ; but that 

 above this it begins to grow rapidly. It is evident, then, that for axial currents greater 

 than half an ampere, we cannot assume a constant permeability throughout the A and B 

 tubes. Even for the a and b tubes, for which one ampere gives an estimated average 

 field of nearly 0*16, it would be unsafe to assume a constant permeability for stronger 

 axial currents than half an ampere. 



It will be seen from the curve that the permeability is almost accurately proportional 

 to the magnetic field from a field of about unity up nearly to its maximum. The curious 

 hump-like character of the earliest part seems difficult to explain as a result of hetero- 

 geneity in the metal. Whatever be its cause, its existence indicates a remarkably rapid 

 increase of permeability when first it begins to increase. I am not aware that any like 

 peculiarity has been noticed in other magnetic experiments. Permeability curves seem 

 always to begin convex to the axis, and not to become concave until the maximum point 

 is being approached. 



In the experiments in which the inductions were measured by ballistic swings, the 

 swings were taken at make and break of the magnetising current which was passed first 

 in one direction then in the other. From the numbers so obtained it was an easy matter 

 to calculate the residual magnetism in terms of the total induced magnetism. With an 

 axial current varying from 0*25 of an ampere to 4 amperes, and a corresponding average 

 field varying from 0'055 to 0"882, it was found that the ratio of the residual magnetism 

 to the total induced magnetism varied from 0'05 to 0'20. 



I now proceed to the discussion of the experiments which form the real subject-matter 

 of this paper. 



The method as already briefly described was a nul-method. Two tubes were taken 

 and subjected to the inductive effect of the same current. This current could be purely 

 axial along both tubes, or purely sectional ; or it could be axial along the one and 

 sectional along the other. To guard against any confusion, I shall here formally define 

 these terms, although the definition is fully implied in what precedes. An Axial Current 

 is a current which flows in a wire forming the axis of the tube, but quite insulated from 

 it. A Sectional Current is a current flowing through the substance or walls of the tube 

 parallel to the axis. 



The following combinations of tubes with currents were tried : — 



B axial 



with A axial. 



B sectional 



)> A „ 



B „ 



„ A sectional. 



B axial 



i) A ,, 



i „ 



„ a axial. 



b sectional 



» * >» 



b sectional with a sectional. 

 b axial „ a „ 



a „ „ A axial. 



a sectional „ A „ 

 a axial „ A sectional. 



