THE AID OF THE ACHROMATIC FRINGES. 99 



method of obtaining unison on the two sides of the bifilar, and as the magnet 

 t' is set by its adjustment screw /', the motion of the fringes while oscilla- 

 ting is seen in the field of the telescope and they are thus never lost. Regula- 

 tion at a, b, figure 70, is more difficult. 



Adjusting the bifilar as to tension in this way, there is one position or 

 distance between h' and t' pretty sharply determinable for which the fringe 

 bands change to stationary ellipses in the absence of all current. This pecu- 

 liar result is at first puzzling, but since it is quite synchronous with the period 

 of the telescope (stationary ellipse), it is obvious that the motion of the 

 objective is the cause of the phenomenon and that the fibers are now in 

 unison with its period. For distances h', t', greater or smaller, the ellipses 

 soon return to bands. The effect of the alternating current on the stationary 

 ellipse is very beautiful. It now oscillates very much like a smoke-ring for one 

 commutator position, whereas it passes in an accentuated way through all 

 phases for the other. Naturally, very complicated displays are also obtained 

 in this double superposition; but practically the vibration of the telescope 

 objective does not disturb the bifilar of the interferometer, unless under the 

 exceptional condition of complete unison, even when both instruments are 

 on the same (insulated) table, a convenience not at all necessary. 



Utilizing the preceding adjustment giving ellipses and bands, respectively, 

 in the two positions of the commutator, many experiments were made to 

 detect a change of phase when a large inductance is placed on both sides of 

 one of the telephones. But in none of the experiments thus far was any 

 difference discernible to be attributed to the presence of the inductance. 

 The endeavor to produce in part, by the mere insertion of inductance, an 

 effect similar to commutation has not, therefore, been realized. 



Among other promiscuous experiments I may refer to the use of a variety 

 of telephones, single and bipolar; to changes in their position, sometimes with 

 their poles (as at mm' , fig. 70) between, and sometimes with the poles on the 

 outside of the wire filaments ee.ee' (as drawn in the figure) ; to bifilars of thread 

 instead of wire; to bifilars of watch-spring; to different sizes and kinds of 

 armatures, etc. Coils, moreover, of different resistances and size of wire 

 were tried, for instance, in figure 73 (which preserves the notation of fig. 70), 

 where C is the coil, h the soft-iron (screw) armature, and M a strong inducing 

 magnet. In the absence of M, no effect was obtained, even when C was pro- 

 vided with a core of soft iron reacting on h. In other words, a magnetizing 

 system is inefficient. In the presence of the magnet M, however, the results 

 were marked, but not better than the above, while the system itself is more 

 complicated. The replacing of h by a hard-steel permanent magnet gave 

 good results, but of inferior sensitiveness. To obviate the annoyances of 

 contact between armature h and M, the latter might advantageously be 

 replaced by a small magnetizing coil surrounding the free end of h and 

 supplied separately with current. Endeavors were also made to utilize the 

 repulsion between a permanent magnet at h and a similar pole at M. But M 

 in such a case reversed the polarity of h. 



