STEEL, NICKEL, AND COBALT TUBES IN THE MAGNETIC EIELD. 463 



the same feature is shown, but the change of sign occurs about Field 270. In Tube III. 

 the form of the graph is the same, but no negative value is reached. In Tube IV., 

 however, this characteristic has disappeared, and the behaviour is, broadly speaking, 

 the same as in Tube VII. Notwithstanding this peculiarity in the sigu of m in the 

 narrow-bored tubes, the calculated values of p! v' come out nearly the same for all. 

 It will be readily seen, both from the curves and from Table IV., that the comparative 

 values of m' v ' follow the same law of change as in the B tubes, approximating in 

 value in the tube of narrowest bore, and gradually drawing apart as the bore increases. 



§ 5. The Coiled Nickel and Cobalt Tubes.— These are formed from sheets of 

 metal, each tube being about 10 inches long and 1 inch diameter. After the long edges 

 had been soldered together to form a hollow tube, the changes of length and the 

 changes of volume of the material in various fields were measured, as already described. 

 The tube was then plugged up with brass discs at both ends, and measurements were 

 made of its changes of bulk. 



From these measurements the dilatations A, /*', v' follow at once. It seems hardly 

 necessary to trouble calculating /x and v in these cases of very thin-walled tubes. A 

 glance at formula (3), § 3, shows that the comparative smallness of V makes the 

 quantities m, v differ very slightly from m', v. 



The three Nickel Tubes C I., C II., C III. were formed from three sheets of 

 different thicknesses. The dimensions are given in Table I. The nickel was obtained 

 as pure as possible, and in this respect is better than the nickel of the bored tubes, 

 which contains 2|- to 3 per cent, of impurities. This may account for the fact that the 

 longitudinal contraction is distinctly greater in the C tubes than in any of the others. 



It is, however, in the external volume changes that the greatest difference is shown 

 between the coiled tubes of very thin wall and the bored tubes of comparatively thick 

 wall. With all of the C tubes there is increase of external volume except in the lowest 

 fields ; in every other instance decrease was the characteristic feature. Compare, for 

 example, the Curves C L, C II, C III. in the lowest left-hand corner of Plate I. with 

 the curves in the first row and with similar curves in the former paper. 



It is interesting to note that the volume increase is greatest in the tube of thinnest 

 wall, and falls off as the wall is made thicker. With a still thicker wall the volume 

 change might change sign and become negative. It would not be safe, however, 

 to institute any strict comparison between tubes formed by coiling sheets and tubes 

 formed from solid bars by boring. 



Excepting that the // curve lies higher than the v curve, there is not any great 

 diversity shown in the nature of the linear dilatations in the various types of tube. 

 The C I. and C III. curves, occupying the middle of the last row in Plate I., are 

 very similar to the A, //, v' curves in the third and fourth rows of the same 

 plate. It may be mentioned that C II. differs so very little from C III. as hardly to 

 require a separate set of curves. The greater divergence between the values of // and 

 v in C I. than in either of the others is noteworthy as being somewhat unexpected. 



