side of the circle is an annulus of wood having an outside diameter of 150 centimeters, an inside diameter of 80 centimeters, and a thickness of 20 centimeters. This float of wood rests on mercury contained in an annular trough of cast iron, of such dimensions as to leave a clearance of about one centimeter around the wood, which space is filled with mercury. It requires about 275 kilograms of mercury to float the entire apparatus which weighs about 1200 kilograms. The float is kept central by a loose-fitting centering pin which sustains no pressure. The annular iron tank is supported by piers of brick or concrete at such a height as to bring the eyepiece of the observing telescope level with the eye of the observer when he takes the posture for easy walking around with the interferometer as it rotates slowly on the mercury. The cast iron trough for the mercury together with the circular wooden float are the same parts as were used in the original Michelson-Morley interferometer of 1887 and these two pieces have been continued in use by the writer to the present time. The other parts of the apparatus of 1887 have been dispersed, excepting only three of the cast iron supports for the mirrors.
The optical flat surfaces were all made in 1902 by that artist-optician, O. L. Petitdidier of Chicago, and proved to be exceptionally perfect; these consist of two plane-parallel plates, each
10.5 × 17.5 centimeters in size, and sixteen plane mirrors of circular shape, 10.25 centimeters in diameter. The general plan of the interferometer is shown in the diagram, Fig. 6, which, however, is not drawn to the exact scale. On a central plate, at the intersection of the arms of the cross, are mounted the half-silvered diagonal mirror, D, and its compensating plate, C, both having been cut from a single plane-parallel disk. At the outer end of each cross-arm, four of the circular mirrors are mounted in a metal plate which is supported in a vertical position. Each of the eighteen mirrors is held by springs against the points of three adjusting screws to permit the necessary adjustments for securing interference. In order to have everything about the two arms as symmetrical as possible, there is no micrometer screw for moving the end mirror parallel to itself, all the adjustment being obtained by means of the three simple screws, as for the other
mirrors. Light from the source, 5, is rendered parallel by the triple-lens condensing system, L, of 15 centimeters diameter, and reaches the half-silvered mirror, D. Part of this light is transmitted to the mirror, I-1; it is successively reflected to mirrors, 2, 3, 4, 5, 6, 7, and 8, having travelled a distance equal to about seven and a half times the length of the arm of the cross. From mirror 8 the light returns by the same path back to D, where it is partially reflected to the observing telescope, T. A second portion of the light incident on D is reflected along the other arm of the cross to II-1, is reflected to and fro and is returned to D and is in part transmitted to the observing telescope. In the actual apparatus, Fig. 10, the mirrors 5 and 7 are above mirrors 3 and 1 instead of at the side of them, and mirrors 6 and 8 are above mirrors 4 and 2. By this system of mirrors the effective length of the arm of the interferometer is greatly increased and in the actual apparatus it is 3203 centimeters, giving a total light-path, going and returning, of 6406 centimeters, equal to about 112,000,000 wave-lengths of the acetylene light used in the experiment. The telescope had an aperture of 3.3 centimeters, a focal length of 35 centimeters, and a magnifying power of thirty-five diameters. The telescope is focussed on the surface of mirror 8
Fig. 6. Plan of the optical paths in the interferometer.
where, when the adjustments are completed, the interference fringes appear to be located.
The apparatus as described, consisting of the optical plane surfaces, the steel cross and the mercury tank and float, has been used by the writer in all experiments from 1904 to the present time, except that for the experiments of December, 1921, the steel cross was replaced with a base of concrete. In 1923, the small reading telescope was replaced by an astronomical telescope of 13 centimeters aperture, having a magnifying power of fifty diameters. The whole path of the light in the apparatus is enclosed; this cover was made of pine wood throughout for the experiments of 1904; in 1905 the cover had glass sides for all arms, thus making the apparatus wholly transparent in the horizontal plane; this arrangement, shown in Figs. 13 and 16, has been used to the present time.
Adjustment of the interferometer
When the mirrors are in position, the distances between them, about 425 centimeters, are compared by means of light wooden rods and the mirrors are adjusted so that the two light-paths, each consisting of eight different portions, are approximately equal. Sodium light from the common laboratory type of sodium lamp is used to establish interference; by observing the visibility maxima of the sodium interference system, the adjustment is made for the center of this system where the white-light fringes may be found. When the apparatus was first assembled on Mount Wilson, the time required for the approximate adjustment of the distances between mirrors with the wood rods was about one hour, for the centering of the mirrors fifteen minutes, for finding the fringes with sodium light thirty minutes, and for finding the fringes with white light forty-five minutes, or two hours and a half for the entire operation. Upon another occasion, the fringes for sodium light were found with ten minutes of searching and the white-light fringes in thirty-five minutes more. The mercury arc and other monochromatic sources have been tried for the preliminary adjustments but the sodium light is preferred because the middle portion of the interference system can be easily located, which corresponds to equal light-paths in the two arms of the
interferometer. White-light fringes were chosen for the observations because they consist of a small group of fringes having a central, sharply defined black fringe which forms a permanent zero reference mark for all readings. Previous to 1924, a small acetylene lamp of the kind used on bicycles, was the source of light, the lamp being carried on a bracket attached to the end of one arm of the interferometer, as shown in Figs. 10 and 13. Such a lamp produces a concentrated, brilliant and very steady light with the minimum production of heat and the lamp itself is very simple and of small weight and it burns for several hours with little attention. For the observations of 1924 and for part of the observations of April, 1925, the source was placed outside of the interferometer room, as is explained later, and a larger lamp of the kind used for automobile headlights, shown in Fig. 14, was used. In April, 1925, the small acetylene lamp was again adopted, now being placed on the top of the cover of the interferometer, over the central axis, as shown in Fig. 16, the light being introduced into the light-path by two mirrors on the end of one arm. This arrangement has been continued to the present. Monochromatic fringes have never been used in the ether-drift observations, though experimental trials have been made, as is described later.
The interference fringes appear to be formed on the surface of the most distant mirror, optically speaking, No. 8, of the series as described. Attached to the supporting frame of this mirror is a small arrow-head of brass, which projects into the field of view, almost in contact with the mirror, forming a fixed fiducial mark for determining the position of the fringe system. Before beginning observations the end mirror, No. 8, on the telescope arm is very carefully adjusted to secure vertical fringes of suitable width. There are two adjustments of the angle of this mirror which will give fringes of the same width but which produce opposite displacements of the fringes for the same change in one of the light-paths. Very careful attention is required always to secure that adjustment of this critical angle which causes the arrow-head pointer to appear to the right of the central black fringe when the light-path of the telescope arm of the interferometer increases in effective length; a
