of Palladium caused by Absorbed Hydrogen. 133 



hydrogen and becomes much crumpled on being charged 

 and discharged. The design of the pole-pieces is given in 

 the next section. 



Several points of the circuit supplying the magnet could 

 be connected at will with a voltmeter, and, by using currents 

 which brought the voltmeter pointer exactly to a division of 

 the scale, the current could be kept accurately constant over 

 a range of 1 to 7 amperes. 



The fine adjustment of the current was made by a weighted 

 wire passing over a groove in a wide brass tube. The wire 

 used for Bovvden brakes was found suitable, and a run of 

 about 10 feet was enough for the adjustment. 



The voltmeter had to be placed at some distance to avoid 

 the influence of the electromagnet on its field, and was read 

 through a field-glass, an arrangement which had the 

 advantage of automatically preventing parallax error in the 

 reading. 



As it was important for the ballistic measurements that 

 the time taken to break the circuit should be constant, a 

 mercury key actuated by a spring was included in the 

 circuit. The core of the electromagnet was earthed to get 

 rid of electrostatic effects. 



3. The Design of Pole-pieces to give a constant 

 Ponder omotive Force. 

 Suppose the field to be symmetrical about the axis of x. 

 Let x, r be cylindrical coordinates. Then, if the pondero- 

 motive force is constant, 



~H. x 'd~HL X fdx = const. 

 H^ = const. X x. 



Again, the flux across any section of a symmetrical tube 

 of force is constant, 



H ;c r ,2 = const, for any line of force. 



Therefore the equation of a line of force will be 

 t i«r 4 = const. 



Let <f> denote the inclination of a line of force to the axis, 

 . . dr r 



then ten + -&=—&■ 



Then, since the lines of force are normals to the equi- 

 potential surfaces, the curvature of the latter at the axis will 



obviouslv be - = ( ^- ) = — -j— , where <f> is now considered 



as a function of x and r. 



