STEEL, NICKEL, AND COBALT TUBES IN THE MAGNETIC FIELD. 459 



inside the brass tube already mentioned, which was adjusted within the magnetizing 

 coil so that the iron or nickel tube was centrally placed in the coil. The mirror rested 

 by two colinear knife-edges on supports firmly attached to a brass cap screwed to the 

 top of the brass tube. A glass rod of suitable length rested by its lower end on the 

 top of the iron or nickel tube, and supported on its upper end a third knife-edge also 

 fixed to the mirror. This knife-edge was parallel to, but lay 1*1 mm. behind, the 

 common line of the other two knife-edges. Any change of length in the inner tube 

 would produce a rise or fall of the support of the single knife-edge, while the two 

 colinear knife-edges would be unaffected. The consequent tilt given to the mirror, 

 which was set approximately vertical, was measured by means of a telescope and 

 reflected scale in the usual manner. A simple calculation gave the corresponding 

 change of length of the tube, and from that the longitudinal dilatation could at once 

 be found. 



Experiments (1) and (2) were made with the large tubes also. In the measurement 

 of the change of length, however, the method was slightly modified. The brass cap 

 supporting the colinear knife-edges was screwed on to the top of the iron, steel, or 

 nickel tube, while the glass rod supporting the single knife-edge passed up through 

 the hollow core from the base of the tube. 



In the third form of experiment the iron or nickel tube was placed within the brass 

 tube, which was otherwise filled with water, and to which the capillary was attached. 

 Any change of volume of the material of the immersed tube produced its effect on the 

 position of the end of the water column in the capillary. 



In the fourth form of experiment the iron or nickel tube was plugged up (air only 

 being inside), and in this condition was dropped into the brass tube, while everything 

 else was exactly as in experiment (3). 



Unless the differences in the mechanical constraints to which any tube was sub- 

 jected in these various experiments produce really important disturbances, we should 

 expect to find the apparent volume change of experiment (4) to be equal to the 

 algebraic sum of the volume changes of experiments (1) and (3). 



A glance at the numbers given in Tables II. and VI. below will show how satisfac- 

 torily the experiments establish this relation. 



§ 3. The Strain Coefficients. — In these experiments the quantities directly 

 measurable are : — 



V, the volume of the material ; 



v, the volume of the bore ; 



v', the volume of the space enclosed by the outer surface and ends — in other 

 words, the volume of the original bar from which the tubes were formed ; 



SV, Bv, oV, the changes of these volumes in given fields ; 



A, the longitudinal dilatation of the tube ; and 



8, = SV/V, the average cubical dilatation of the material of the tube in these same 

 fields. 



