STEEL, NICKEL, AND COBALT TUBES IN THE MAGNETIC FIELD. 4G1 



This process has been applied to the old iron tubes 1. 3, 5, VII. , to the new iron 

 tubes T. to Vlir., and to tube A VII. 



As will be seen immediately, the cubical dilatation in the case of nickel is negligibly 

 small compared with the elongations, so that \il is to be found from the simplified form 

 of equation 



A + 2m , = ^(X + 2m) (3)' 



The ratios v and v follow as above. 



The values for the Nickel Tubes 1, 4, 7, and I. to VII. have been obtained in this 

 way, the additional assumption being made that the elongation X is the same for the 

 non-existent Tubes I. to VI. as for the final existent form VII. A consideration of all 

 the measured values of A for the various tubes, large and small, will show that this 

 assumption, though not strictly true, does not involve an error of magnitude sufficient 

 to modify seriously the final conclusions. 



§ 4. The Bored Nickel Tubes. — These are best considered first, because of the 

 comparative simplicity of the results obtained. The volume changes and dilatations 

 of the B tubes in various magnetic fields are given in Table II., and are shown graphi- 

 cally in the first two rows of Plate I. 



Especially noteworthy is the smallness of the material volume change in comparison 

 with the other measured volume changes. So minute is SV, that in calculating the 

 strain coefficients we may, without any risk of serious error, put 



The Bv and Sv' curves lie very close together. It is, in fact, hardly possible, on the 

 chosen scale, to draw them distinct in the case of B II. For B V. one curve only is given, 

 that, namely, which shows how the ajjparent external volume change increases with the 

 field. A tiny crack in the wall of the tube prevented any good observation of the 

 bore change being made. 



On the whole, there is a tendency for the volume changes Sv and Sv' to differ more 

 as the bore increases, that is, as the walls get thinner. This may be referred to the 

 different conditions of constraint in the two forms of experiment. When Sv was being- 

 measured, the brass cap, by means of which the capillary was attached to the upper end 

 of the tube, was screwed on to the outside surface. On the other hand, in order to per- 

 mit the tube to slide easily within the brass tube when Sv' was to be measured, the nickel 

 tube was in this case closed by a brass plug which screwed on to the inside surface. 

 Thus in any lateral expansion of the tube, there would be more constraint with the 

 outside fitting cap than with the inside fitting plug. 



Again, the manner in which the cubical dilatation diminishes with the thickness of 

 the wall suggests the possibility that part of the measured change of volume SV 

 may be due to empty spaces within the metal — in other words, to its vesicular structure. 



Passing to the consideration of the coefficients of strain, we notice a steady, though 



