|283 284 285 286 287 288 289 290 291 292 293 294 295 296|
such interest that the section of the Association before which the paper was read, obtained, from the research funds of the Association, a grant of money with which to construct the apparatus and make the experiment. The apparatus was finished in the summer of 1890, and Professor Eddy came to Cleveland to take part in the experiment.
The optical part of the apparatus consists of the interferential re fractometer used by Michelson in his experiments to determine
whether light moves with the same velocity in different directions in the solar system. A train of waves of light from the source s, (Fig. 1) is divided at the second surface of aby which is coated with silver of such thickness that the division is nearly equal. If the mirrors, / and gy are properly adjusted, the two parts reunite at ab} and interference bands can be seen by the eye at e, if the distances off and g from ab are properly related. If, now, any force acts intermittently to change the velocity of one part of the ray while the paths are separated it may be detected by a motion of the interference bands, provided the change is not too small.
If we put two tubes, h and z, in the paths of the divided ray, make these of very nearly the same length, close them with plane parallel glasses of equal thickness, and fill the tubes with carbon bi
sulphide, we shall still obtain interference bands. If we surround each tube with a coil of copper wire, use polarized light, and close the circuit through one coil, there will be a rotation of the plane of polarization of the ray which passes through the magnetic field. If the rotation be not so much as to prevent, or too much affect, the interference of the parts of the divided train of waves, a change of velocity of light while in the magnetic field will produce a displacement of the interference bands. If we complete the circuit in the two coils alternately, we double the displacement which is sought to detect.
If any displacement is observed, it will be necessary to determine whether it may not be due to strain or displacement of parts of the apparatus. In the case of the plates ab, cd, f and g, this may be accomplished by watching the effect of closing alternately the circuits through the two coils while tubes i and h contain only air. But a disturbance which should affect the position or length of one of these tubes needs special means for its detection. The plates which close the ends of these tubes are accordingly placed as shown at k, /, m and 11, and are silvered except where they cover the tubes. When proper adjustments are made, the observer can see interference bands through the tube by placing the eye at c, interference bands at the covers of the nearer end of the tubes by placing the eye at c\ and interferences at the covers of the farther ends of the tubes by placing the eye at cn. Since the interferences produced in these three cases are due to rays which have passed through only glass and air, they can be so adjusted as to detect with ease and certainty a displacement which would be entirely insensible in the irregular and most unstable interference bands produced after a ray of light has passed through a liquid so optically unstable as carbon bisulphide. Our means of detecting the effect of mechanical disturbances were therefore greatly superior to those of detecting the change of velocity in question, and the statement that no displacement of interference bands was caused by mechanical disturbance due to the opening and closing of the current through the coils is open to no doubt.
The plates ab and cd were 9.3 by 2.0 cm.; / and g were 2.0 cm. by 2.0 cm.; k, /, m and n were 4.3 cm. by 2.0 cm. The tubes Ji