THE AID OF THE ACHROMATIC FRINGES. 73 



deflection 5 was increased 20 or 30 per cent. If we suppose the more marked 

 induction occurs on breaking the primary circuit, the equatorial currents 

 induced in the thick brass tube would coincide in sense with the magnetization 

 and this would be made obvious by an increase of s. In a repetition of this 

 experiment some time afterward the brass tube behaved exactly like the iron 

 tube, the deflection being diminished over 50 per cent in both cases. 



Finally, with the solenoid so placed that there was no displacement of fringes 

 (normal bands), both the iron and the brass tubes or screens produced fringe 

 displacement, the iron without change of phase (oblique bands), the brass 

 with change of phase (ellipses). 



52. Amplitude of the interrupter. One curious result, however, was ob- 

 tained by inserting additional resistances R=i to 10 ohms in the primary 

 circuit. It was found that the secondary current passes through a maximum 

 and does not as a rule decrease with the resistance R added to the primary. 

 Naturally the band-width, which marks the amplitude of the objective, does 

 regularly decrease with R. An example may be given (ring transformer) 

 the ohmic resistance of the circuit being below i ohm: 



R= 012345 10 ohms. 

 5=10 10 12 18 15 ii 7 



With the solenoid similar results were found, for instance (internal ohmic 

 resistance i . 7 ohms) : 



R= o i 2 3 5 10 ohms. 



5=13 16 18 15 12 5 



The maxima here are sharp, almost cuspidal. Hence the only explanation 

 which occurs to me is a consideration of the resonance of the interrupter of 

 the primary circuit and the vibrator in the secondary. As the interrupter 

 plays through insulating water into the mercury contact, its period will 

 probably increase with its amplitude, because of the damping effect of the 

 water. The sharp maximum is determined by coincident periods at the 

 interruptor and the vibrator. 



53. Telephones in parallel. The telephones actuating the vibrator were, 

 in the above work, connected in series, because they could then be used either 

 in concert or in opposition, as indicated. There may, however, be an advant- 

 age in joining them in parallel and excluding the auxiliary telephone, when a 

 low resistance is needed in the secondary. This would fall to about one-fourth 

 and thus be below 50 ohms. I made a number of experiments in this way with 

 results essentially like the above. In case of the ring transformer and a single 

 turn of secondary, the ellipses remained in the field without external resistances 

 CR = o). Under these circumstances it was possible to exchange an inductive 

 and a non-inductive resistance allowing for the difference of phase. The 

 sensitiveness obtained was about 5Xio~ 8 ampere scale-part. 



