PHYSICS: C. BARUS 335 
5. Decreased Bifilar Distances. — The easy accomplishment of the 
above experiments, where the distance between the bifilar threads was 
about 30 cm. and their length 60 cm., encouraged me to reduce the dis- 
tances between the threads until the system was virtually torsional. In 
such a case the displacement at the ends of the vibrating beam {cc\ 
figure 4) is no longer Umited to that of the bifilar wires. Figure 4 is 
a front elevation of the new apparatus; the steel wires dd' being at a 
distance of about 6 cm., their length 60 cm., the distance between the 
auxiliary mirrors m, m\ 10 cm., and between the magnets of the tele- 
phones T, T' , 16 cm. The vibrator cc' was (as above) a steel umbrella 
rib. The ends of the wires dd' were again wound about the stiff screws 
h, h' held in posts a, a\ rigidly attached to the bed plate of the interfer- 
ometer (not shown). The wires were stretched below by the pulley w 
and screw y, controlled by the nut u pressing against the rail also 
rigidly attached to the bed plate of the interferometer. Provision must 
be made (slotted sheath z and pin p) to prevent y from turning on its 
axis. A variety of braces are to be introduced to obviate synchronous 
vibrations of parts of the apparatus. 
The adjustment for resonance is here more difiicult than in the pre- 
ceding case, and y must be a fine or a differential screw and the nut u 
work smoothly. Resonance should first be established by aid of the slit 
image in a telescope with fixed objective and with small resistance 
(200 ohms) in circuit. After this a large resistance (10,000 ohms and 
more) may be inserted and the work continued with the interference 
fringes when further adjustment is possible. 
In a later design of the apparatus, the bar v, rigidly fixed at one end, 
was elastically controlled at the other end by the fine threaded screw t, 
pushing against the rigid socket s. This is an excellent fine adjustment 
and when used does not disturb the fringes. 
6. Observations with the New Apparatus. — The first trials were made 
with the white slit image. Using the coil with a virtual electromotive 
force of about 0.7 volts and a small resistance (200 ohms), the resonance 
conditions were easily reached in the well-braced apparatus. The 
resistance was then increased to 10,000 ohms and the corresponding 
band- width found. When obtained the band-widths s, persists. The 
sensitiveness observed if is 18 scale parts, was 0.7/18 XlO-* = 4XlO-^ 
amperes per scale-part (0.01 cm.) of the sht image. Per A iV = 10~* 
cm. of the micrometer of the interferometer, where hd = AN cos i and 
6 = s/2 f (f = 23 cm. being the focal length of the telescope, b = 10 
cm. the breadth of the ray parallelogram, i = 45°), the current would 
be estimated as 1.3 X 10~^ ampere. 
