Effect of a Changing Magnetic Field. 



467 



From this the deflexion due to the electromagnetic effect 

 should be about *1 mm., or the difference between deflexions 

 in opposite directions '2 mm. No such small difference could 

 with certainty be determined, as the variations of individual 

 readings were generally much larger*. Although slight 

 improvements could probably be made and the magnitude of 

 the deflexions somewhat increased, still the present method 

 seemed incapable of yielding a decisive result, and a new 

 and considerably different plan was adopted. 



Fig-. 2. 



Second Method. 



The form of apparatus which finally proved successful is 

 constructed as follows: — One of the conductors of a parallel 

 plate condenser is built up of 20 rings of 

 thin varnished cardboard, inside diameter 

 10 cm., outside diameter 18 cm. (a, a, fig. 2). 

 A strip 2 cm. in width along the inner edge 

 of each ring is covered with silver leaf. 

 The rings are fastened parallel to one 

 another and about 9 mm. apart by a light 

 wooden frame (b, />), and the frame is sus- 

 pended by Hne wire and phosphor-bronze 

 ribbon (s, s) fastened above and below to 

 torsion-heads and stretched tight, the rings 

 themselves lying horizontal. The other 

 portion of the condenser is stationary and 

 consists of 21 rings of thin hard rubber 

 (p, p), inside diameter 8 cm., outside dia- 

 meter 14 cm., placed so as to alternate with 

 the suspended rings. A 2 cm. strip along 

 the outer edge of the hard rubber rings 

 is covered on both sides with tinfoil, and 

 the rings themselves are held in place by a 

 hard rubber tube (g, q) of 8 cm. outside 

 diameter. Through the centre of this tube, 

 and not touching it, passes an iron core 

 (r, r) about (3 cm. in diameter, built up of 

 fine shellaced wire. This core forms part 

 of a closed rectangular magnetic circuit 



* On one occasion three series of readings gave remarkably consistent 

 results, each series consisting of from 12 to 20 pairs of readings. How- 

 ever, the observed ditierences were very considerably larger than would 

 be demanded, by theory, and the results were never duplicated. The 

 whole apparent consistence of results can, therefore, have been due only 

 to a series of coincidences. 



2 112 



