42 



DISPLACEMENT INTERFEROMETRY BY 



around a, downward by the groove to D and then up in the corresponding 

 way to b f . The lens carriage gg' is then attached with cement (after the wires 

 are evenly stretched) with the cross-wires r, s on the opposite side of dd' to 

 the pull of the magnet i. The magnet, in addition to the cement, thus guards 

 against slipping. On turning the screws b and b', any degree of tension may 

 be imparted to the wires dd', roughly. This simple device worked surprisingly 

 well, and wires of different kinds may be easily inserted or replaced, the lens 

 system being subsequently attached with cement ; but it is better to loop the 

 lower part of dd' around a special roller G, as indicated in figure 42 and used 

 in my later tests, with the object of more easily reaching an equality of tension 

 in the wires d and d'. The tensions are then roughly changed by the screw 

 and nut u. 



e/ 



if I? 



The approximate tension having thus been obtained by the screws 6, b', 

 or u, the finer variations are imparted by the screw mn which flexes the elastic 

 rectangle ABCDE and thus gives to the bifilars dd' exactly the tension 

 required. It is at the thumb-nut m that all adjustment is made during 

 observation. 



In my apparatus the rectangle was about 50 cm. long and 12 cm. wide. The 

 wires dd' about 1.5 cm. apart and each about 0.025 i 11 - m diameter. Wires 

 as thick as this require sharp adjustment as to tension, but they obviate, 

 when tense, all objectionable quiver and the method given proved quite 

 satisfactory. The tension is sufficient to admit of an air-gap of less than a 

 millimeter between the plate g' and the magnet i of the telephone. Later, 

 in the interest of sensitiveness, the telephone hi was also put on a stout spring 

 micrometer-screw system so that the distance ig' could be nicely regulated. 



As a source of light a distant vertical electric filament or a slit and collimator 



