68 DISPLACEMENT INTERFEROMETRY BY 



equal effects on telephone and vibrator. In another experiment the closed 

 circuit gave 20 scale-parts. The insertion of 4 microfarads decreased this to 

 4 scale-parts, which is again a demand of about 20 microfarads for an equality 

 of behavior. On the other hand, while the telephone responds for a phenom- 

 enally small capacity, it soon ceases to increase in loudness (for i, 2, 3, 4 mf. 

 or even 250 ohms), whereas the deflections of the vibrator increase regularly. 



48. The vibratory stray magnetic fields. As the narrowed bifilar of 47 

 had virtually failed, I returned to the apparatus of 42, with bifilar distance 

 of about 6 cm., this being the more serviceable. The vibrating needle was 

 here a steel wire with adjustable mirrors on bits of cork. The fringes were 

 easily found. Instead of the telephonic inductor of 46 (which, though inter- 

 esting in itself, gives a tumultuous effect on the reversal of the magnetic 

 field), a simple spring commutator (old Ruhmkorff style) was put in place 

 of the key K (fig. 59), receiving the poles of the secondary. In such a case 

 any secondary could be used at 5, and its alternating currents either reversed 

 or broken at pleasure. This proved to be an excellent method, for the ellipses 

 now merely glided into new positions without passing through oscillations. 

 The secondary system was, as it were, dead beat, though still somewhat 

 slower than would have been desired in the interest of expeditious work. 



Before proceeding further it was necessary to learn to control the so-called 

 mechanical coupling. The probable reason has already been suggested in 

 45. By placing the primary adjustably on a table at a distance of about a 

 meter or more from the vibrator, the fact that the latter is subject to all the 

 quivering stray magnetic lines in the room was soon verified. I may recall 

 that the primary coil consisted of a thin (walls 0.08 cm.) iron tube, 55 cm. 

 long, 0.635 cm. in diameter outside, wrapped with a single layer of wire 

 (0.034 cm. in diameter), so that about 21 turns came to the linear centimeter. 

 On this stick-like solenoid any secondary could be slid at pleasure. The 

 primary was now fastened in a horizontal position to a standard, in a way 

 admitting of rotation around a vertical axis into any fixed position. With 

 the secondary broken (R = ) it was possible to obtain any phase or ampli- 

 tude of ellipse within limits by rotating the primary solenoid. Thus either of 

 the cases, figures 64 and 65, for R= < could be produced at will. If the 

 solenoid is too far off, etc., the figure of minimum amplitude may be a sym- 

 metrical ellipse, so that the bands parallel to the vibration line of the objective 

 are unattainable. Otherwise it is as easy to compensate the vibrator as it 

 would be, for instance, to compensate a galvanometer by an astasizing magnet. 



When this compensation has been made the vibration figures obtained by 

 switching the secondary current at K are in opposite phases and the measure- 

 ment of the range of ellipses may be made with the same result in either phase. 

 It is also possible to change the phase of ellipses produced by a secondary 

 current, obtaining a series of oblique bands corresponding to the current. 



The amplitude of the ellipses for the case of direct magnetic action (second- 

 ary current absent, R= ) depends both on the strength of the alternating 



