58 DISPLACEMENT INTERFEROMETRY APPLIED TO 



51. Shadow waves. Oscillation about horizontal axis. It seems improb- 

 able, at the outset, that any vibration figure of the type 84 should be static on 

 the telephone plate. One would preferably infer that the form rotates about an 

 axis normal to the plate. True, in the bipolar telephone there would not be 

 concentric symmetry of the magnetic induction, but the old unipolar behaves 

 essentially like it. The phenomena, as a whole, occur in all telephones which 

 I have examined, and they are not, in general, liable to be accidental. I was, 

 therefore, at first inclined to suppose that the telephone plate vibrates with 

 different periods which interfere, something after the manner of a bell. Later, 

 however, it developed that the plate, firmly clamped in a hard-rubber case, not 

 rigorously circular and very subject to changes of temperature, must be the seat 

 of static strains (buckling), varying in the course of time as the room is hotter 

 or colder. 



It is easy to show that the vigorously energized telephone plate oscillates 

 round a horizontal axis in the plate, also, by merely drawing a very fine wire 

 across the slit of the collimator. This produces a dark shadow line in the 

 bright field of the vibration telescope and the line frequently takes the form 

 of the beating wave- trains when the amplitude of the telephone plate is 

 increased. Moreover, the maxima of grouped shadow waves vibrate in a 

 manner similar and with the same period as the lines s, and are frequently 

 associated with them in the way suggested in figure 82. In other words, the 

 periods of oscillation round horizontal and vertical axes in the plate is nearly 

 but not necessarily quite the same. 



It follows from this that the axis of oscillation in the plate is generally 

 oblique. If not, either the lines or the shadow waves fail to appear, as is often 

 the case. I could not, however, ascertain whether the axis rotates; for the 

 amplitudes, etc., of the shadow waves are not sufficient for this discrim- 

 ination. Similarly, if the vibration telescope is at rest, the coincident slit- 

 images do not usually separate (which would destroy the fringes so that gaps, 

 such as are mentioned below, occur in the fringe band), but the duplicated slit- 

 image merely broadens. The angular displacement in either case would 

 probably not exceed io- 4 radian (within 0.01). The corresponding angular 

 displacement of the fringes, however, may be several hundred times greater. 



52. Beating fringe waves. The reason for the synchronous behavior of the 

 fringe waves and slit-images is thus at hand. When these are very slightly 

 separated laterally, the fringes move in a marked manner up or down. Again, 

 if the slit-images are slid slightly along each other longitudinally, the fringes 

 also move up or down, supposing, of course, that in both cases the trend 

 of fringes at rest is horizontal. Consequently if the horizontal and vertical 

 periods are not quite the same, since both produce directionally the same 

 displacement of fringes, there must be two beating wave-trains in the field of 

 the vibration telescope, quite apart from other and more immediate reasons. 



This is, moreover, in keeping with the long periods of the vibrating groups as 

 given in the table, and with their changes in the lapse of time, or in different 



