134 Mr. H. F. Biggs on Decrease in the Paramagnetism 



If, then, the faces of the pole-pieces are spherically curved! 

 with radii of curvature differing by four times the distance 

 between the centres of the faces, these faces should coincide- 

 nearly enough with sucli equipotential surfaces, and should 

 give a field with approximately constant ponderomotive force 

 for points on or near the axis. Further, the radii r v r 2 of 

 the faces (measured perpendicular to the axis) should 

 correspond to the radii of a single symmetrical tube of force- 

 Thus we should make 7 > 1 /r 2 =(« 2 / / »i) 1/4 . In the preliminary 

 experiment referred to, the palladium was balanced against 

 weighed bits of bismuth in a field thus designed. The 

 apparatus was very simple, and no knowledge of the field 

 was necessary. The method gave results correct to about 

 10 per cent. In the final experiment larger pole-pieces- 

 were used, whose dimensions were those given in fig. 1. 



Fig. 1. 



#x=2 cm. 



pi= 8 cm. 



r l = 4:'5 cm 



x ± — o cm. 



p 2 = 20 cm. 



r 2 = 3'5 cm 



The edge of the concave pole-piece was rounded as shown 

 to avoid disturbance of the field by sharp edges. To test 

 the uniformity of the force a glass rider was made with a 

 bit of bismuth at one end and a bit of aluminium at the 

 other. The rider was placed on the coil in such a way that 

 the ends lay in the axis of the field, and could be turned 

 either with the bismuth in the strong field and tho aluminium 

 in the weak field or vice versa. By noting the difference of 

 the balancing-current in the coil, when the rider was- 

 reversed, it was found that the ponderomotfve force only 

 varied by about 1 per cent, over 1*3 cm., the distance 

 between the ends of the rider. 



