STEEL, NICKEL, AND COBALT TUBES IN THE MAGNETIC FIELD. 469 



when the iron tube became magnetized, extra pressures between it and the bottom 

 or walls of the brass tube might very easily come into existence ; and these might 

 reasonably enough be supposed to produce minute but appreciable changes of volume. 

 In one experiment the tube was set 2 inches above the usual position in the 

 magnetizing coil ; but this displacement from the central position did not to any 

 decisive extent affect the readings — a point which rather tells against the explanation 

 just given. But the most serious difficulty is the magnitude of the strain, which, 

 though small, represents stresses of considerable magnitude. The micrometer-reading 

 14 means a volume change of 34 x 10~ 6 cub. cm. The bore of the brass tube was 

 46 cms. long and 3 cms. in diameter, giving a cubical dilatation of fully 10~ 7 within the 

 tube. We may get an approximate idea of the pressure required to produce this 

 dilatation by solving the equation 



Pa 2 - P'6 2 1 



io- 7 = 



b 2 - a 2 k 



where P P' are the internal and external pressures, a and b the inner and outer 

 radii of the tube, and h the reciprocal of the compressibility. In the present case b is 

 2*1 cms., h may be put equal to 10 12 C.G.S. units, and P' is the atmospheric pressure, 

 say 10 6 . Thus— 



Pa 2 = 10 6 (4-41 + -22) 

 and 



P = 10 6 x 2-06. 



Hence it would need an additional pressure within of fully an atmosphere to produce 

 the dilatation observed. This seems hardly possible under the conditions, for the 

 interior communicates with the outside through the capillary. 



The extraordinary manner in which the plugged Tube A III. passed through 

 its changes when the field was removed was common to the tubes from A I. to 

 A IV., and from B I. to B III. But there was no evidence of it with A V., B IV., and 

 the tubes of wider bore than these. In all cases in which it occurred, the effect was 

 the same — an increased negative change at the instant the magnetizing current was 

 broken, followed immediately by a return to a final positive or diminished negative 

 change. The phenomenon may be explained as an effect of the current of self-induction, 

 producing a short-lived but intense magnetization in the outer layers of the iron tube, 

 for the phenomenon is not observed in the experiments for measuring the change of 

 volume of bore — that is, in experiments on the behaviour of the tube at the inner wall. 

 If this, however, were the sole cause, it is hardly likely that it would so entirely 

 disappear in A V. and B IV., which still have fairly thick walls. It is conceivable 

 that there may also be a tendency for the transverse dilatation /*' to disappear more 

 quickly than the longitudinal dilatation X. A slight tendency in A to persist when the 

 field was removed — in other words, an appreciable time-lag — would, in the higher fields, 

 give rise to a momentary increase in the diminution of volume. 



