98 



DISPLACEMENT INTERFEROMETRY BY 



the commutator positions I and II and resistances R in circuit, the results 

 were, for instance : 



The reason of this is apparent, for when the telephones are joined in 

 series the mirror mm' (fig. 70) is periodically rotated and released around a 

 vertical axis and the displacement of fringes is proportional to the small 

 angular amplitude of rotation. If, however, the telephones are connected 

 differentially, the mirror mm', if properly adjusted, merely moves parallel to 

 itself, fore and aft, and the fringes remain stationary. More usually, how- 

 ever, there is a difference in the size of ellipses (cat. par.} in the two cases and 

 at other times there is scarcely any difference at all appreciable. In such cases 

 it seems probable that the periods of the two filaments ee and e'e' on opposite 

 sides of mm' are not the same, or differ in amplitude (attracting forces of 

 different strengths) , so that the fore-and-aft motion is accompanied by more 

 or less rotation, residually. It is in fact difficult to make the two telephones, 

 etc., quite identical in action. 



It is for the investigation of this question that the adjustment pushing- 

 screws /, /' and springs 5,5' (pulling toward the rear), the telephones being 

 on a vertical axes, were provided. If one of these, / for instance, is placed 

 in a definite effective position, while t' is relatively far from its armature h f , 

 the screw /' may be gradually pushed forward, diminishing the distance to 

 the minimum. The effect of this is further to rotate mm', and if the turning 

 of /' is cautiously done the fringes may be passed from top to bottom of the 

 telescopic field by /' and restored to position by the micrometer. After each 

 step of the experiment the fringes are observed, when actuated by the 

 alternator, in the two positions of the commutator of one telephone. In this 

 way it is possible to find the adjustment in which for one position of the com- 

 mutator there is excessive motion of fringes, whereas for the other there is 

 practically no motion, as in the example above. If the distance (t',h') is 

 markedly larger or smaller, the distinction of the two positions of the com- 

 mutator is lessened and may even vanish. 



There is usually some vibration figure (i/i, 3/4, etc.) best adapted for the 

 given adjustment, even if the other figures appear. When this is chosen, the 

 distance from magnet to armature (t f to h') makes little difference within 

 reasonable distances. Such a result is not unexpected, for the whole phenom- 

 enon is relative, larger differences corresponding to larger total forces. The 

 distance between magnet and armature (here on one side) does, however, 

 affect the tension of the string, since the forces and the stretch are greater 

 for smaller distances and the period therefore smaller. This is the simplest 



