THE AID OF THE ACHROMATIC FRINGES. 57 



The remarks made presuppose that the telephones are acting in concert, on 

 opposite sides of the vibrator. We may refer to this as an arrangement in series. 

 When the telephones are acting in opposition, the band-width decreases to a 

 few scale-parts, 2 or 3, depending on the symmetry of adjustment, etc., which, 

 if perfectly made, should throw the differential 5 out entirely. 



A large number of experiments were carried out with soft-iron armatures, 

 cemented to the vibrator cc' in front of the telephones ; but no result of conse- 

 quence w r as obtained in this way, so that I returned to the simpler unarmed 

 vibrator. Care must be taken to obviate vibration about the axis of cc', 

 which betrays itself by producing wave-lines across the band-width for every 

 obstruction in the slit. These would be a serious annoyance were the inter- 

 ferometer used. 



41. Observations with the interferometer. As has been stated, the fringes 

 are easily found, because the rapid motion of the vibrator implies considerable 

 damping. The slit-image is thus quite stationary and the fringes clear and 

 strong. On starting the inductor, the fringes at once vanish and after breaking 

 circuit only slowly reappear, unless the vibration telescope is used. If the 

 period of the latter differs from that of the induction (to be very weak) , the 

 even band of fringes, a, figure 60, changes 

 to wave-lines b traveling in opposed direc- 



tions and of continually increasing ampli- 

 tude. Eventually the crests or troughs 

 only are seen (and these but on one side 

 c, figure 60) , as some micrometer adjust- 

 ment for one or the other will be necessary 

 to keep them in the field) again traveling 



in pairs, in opposite directions, through each other. On breaking the circuit these 

 pulses slowly coalesce into the wave-band 6 and finally into the even band a. 



The case here presented is that of a relatively slowly vibrating telescopic 

 objective at V. If the frequency of the latter can be counted, the frequency 

 of the alternator may be deduced from the number of moving crests in the 

 field. Thus, when the frequency of the objective was n'= 5, there were four 

 crests in motion, implying a frequency of n = 20 for the interrupter of 

 the coil. In this respect the case of different periods is advantageous. 



The fringes in these preliminary experiments were chosen small. It is of 

 interest, nevertheless, to compute the sensitiveness reached. The effective 

 voltage of the coil was about 0.7 volt; the resistance inserted 42,000 ohms. 

 Hence the mean current was 0.7/4.2 Xio 4 = io~ 4 /6 ampere. The net double 

 amplitude of the waves measured by the ocular micrometer was about 4 

 scale-parts, so that * = 4Xio- 6 ampere conies to a scale-part in case of the 

 present small fringes. 



The next step in advance consisted in adding an electromagnet (e f , fig. 59) 

 at the objective of the vibration telescope V, in series with the electromagnet 

 e of the interrupter. The two springs at V and /, moreover, were adjusted 



