756 



BELL SYSTEM TECHNICAL JOURNAL 



"arms" of the machine, at the ends of which they meet fully-reflecting 

 surfaces A and B set normal to their courses. Reflected straight back 

 along the arms, they are once more divided by the semi-transparent 



Fig.'_4. 



mirror, and the fractions which go towards the observer (left, in 

 Fig. 4) are those which interfere. This seems complicated; but in 

 essence it is not so. Everything happens as though the mirrors A 

 and B were at the end of the same arm. Looking from the left 

 through C, one perceives the mirror A and the virtual image in C of 

 the mirror B; and this comes to the same as though C could be 

 removed, and the horizontal arm of the machine swung into coincidence 

 with the vertical arm in spite of the well-known prohibition against 

 two objects occupying the same place at the same time. When the 

 plane of A is accurately normal to the plane of B and the semi- 

 transparent surface C is properly oblique, the virtual image of B is 

 accurately parallel to A ; and the observer sees circular fringes, which 

 one by one emerge from a central spot or shrink down into it as either 

 mirror is displaced along its arm. 



This clever device of combining one real mirror with the virtual 

 image of another, to form a thin plate of which one surface is im- 

 palpable, is of enormous value. One can make the virtual reflector 

 go right through the real one, the fringes being swallowed up into the 

 centre one after another, and then reborn in due order after the whole 

 field of view goes black at the instant of coincidence. By moving one 



