June 19, 1925]
SCIENCE
619
of wood was used in 1902 and 1903, but the changes in the wooden frame due to the variations in humidity and temperature made accurate observations impossible. A new supporting frame was designed by Professor F. H. Neff, of the department of civil engineering of Case School of Applied Science, the purpose being to secure both symmetry and rigidity. This frame, or base, was constructed of structural steel, and was erected in a basement room in the Physical Laboratory of Case School of Applied Science in Cleveland, and observations were made in 1904 and 1905. The results of these observations were published in The Philosophical Magazine for May, 1905. They were stated as follows: “We may therefore declare that the experiment shows that if the ether near the apparatus did not move with it; the difference in velocity was less than 3.5 kilometers per second unless the effect on the materials annulled the effect sought. Some have thought that this experiment only proves that the ether in a certain basement room is carried along with it. We desire therefore to place the apparatus on a hill to see if an effect can be there detected.”3
It was at this time that Einstein became interested; and later in the year, 1905, he published a paper on “The electrodynamics of moving bodies.”4 This paper was the first of a long series of papers and treatises by Einstein and others which has developed into the present theory of relativity. In this first paper, Einstein states the principle of the constancy of the velocity of light, tending to show that for an observer on the moving earth, the measured velocity of light would be constant, regardless of the direction or amount of the earth’s motion. The whole theory was related to physical phenomena, largely on the assumption that the ether-drift experiments had given a definite and exact null result. This interpretation of the experiment was not acceptable to the writer, and further observations were undertaken to determine this particular question.
In the autumn of 1905, Morley and Miller removed the interferometer from the laboratory basement to a site on Euclid Heights, Cleveland, at an altitude of about 300 feet above Lake Erie, and free from obstruction of buildings. Five sets of observations were made in 1905-1906, which give a definite positive effect of about 1/10 of the then “expected” drift. There was a suspicion that this might be due to a temperature effect, though there was no direct evidence of this. A plan was made for putting this
3 Morley and Miller: ‘‘ An experiment to detect the FitzGerald-Lorentz effect, ” Phil. Mag., 9, 680 (1905); Proc. Am. Acad. Arts and Sci., 41, 321 (1905) ; “On the theory of experiments to detect aberrations of the second degree,7 ’ Phil. Mag., 9, 669 (1905).
4 Einstein: 11 Zur electrodynamik bewegter Körper, Ann. der Physilc, 17, 891 (1905).
surmise to the test after a summer’s vacation. We had erected the interferometer on land owned by a friend; during our vacation absence, the land was sold and the new owner ordered the immediate removal of the interferometer.
Professor Morley retired from active work in 1906 and it devolved upon the present writer to continue the experiments. It seemed desirable that further observations should be carried out at a much higher altitude, but numerous causes prevented the resumption of observations. The publication of reports on the solar eclipse of 1919, which were interpreted as confirming the theory of relativity, revived the interest in the ether-drift experiments. A generous friend provided ample funds to cover the considerable expense involved. The site of the Mount Wilson Observatory near Pasadena, California, at an elevation of about 6,000 feet, seemed to be a suitable place for further trials. Through the kindness of President Merriam, of the Carnegie Institution at Washington, and of Directors Hale and Adams, the experiments were resumed by the writer in March and April, 1921, at the Mount Wilson Observatory. The apparatus was substantially the same as that used by Morley and Miller in 1904, 1905 and 1906. Observations were also made in the latter part of the year 1921 and again in 1924 and 1925.
At the Mount Wilson station, about 5,000 single measures of the ether-drift have been made at various times of the day and night. These have been reduced in 204 different sets, each set consisting of observations made within one hour’s time. The observations correspond to four different epochs of the year, as follows: I. April 15, 1921, 117 sets of observations; II. December 8, 1921, 42 sets; September 5, 1924, 10 sets; and April 1, 1925, 35 sets.
The very first observations made in March, 1921, gave a positive effect such as would be produced by a real ether drift, corresponding to a relative motion of the earth and ether of about ten kilometers per second. But before announcing such a result it seemed necessary to study every possible cause which might produce a displacement of fringes similar to that caused by the ether drift. The causes suggested were magnetic deformation of the steel frame of the interferometer and the effects of radiant heat. In order to eliminate the effects of radiant heat the metal parts of the interferometer were completely covered by cork about one inch thick. Fifty sets of observations were made under these conditions, showing the periodic displacement of the fringes due to the drift agreeing with the first observations.5
5 Miller: "Ether-drift experiments at Mount Wilson Observatory," Phys. Rev., 19, 407 (1922); Science, 55, 496 (1922).
620
SCIENCE
[Vol. LXI, No. 1590
In the summer of 1921 the steel frame of the interferometer was dismounted and a base of one piece of concrete reinforced with brass was cast in place of aluminum or brass, thus the entire apparatus was on the mercury float. All the metal parts were made free from magnetic effects and the possible effects due to heat were much reduced. In December, 1921, 42 sets of observations were made with the non-magnetic interferometer. These show a positive effect as of an ether drift which is entirely consistent with the observations of April, 1921. Many variations of incidental conditions were tried at this epoch. Observations were made with rotations of the interferometer clockwise and counter clockwise, with a rapid rotation and a very slow rotation, with the interferometer extremely out of level, due to the loading of the float on one side. Many variations of procedure in observing and recording were tried. The results of the observations were not affected by any of these changes.
The entire apparatus was returned to the laboratory in Cleveland. During the years 1922 and 1923, many trials were made under various conditions which could be controlled and with many modifications of the arrangements of parts in the apparatus. An arrangement of prisms and mirrors was made so that the source of light could be placed outside of the observing room, and a further complication of mirrors was tried for observing the fringes from a stationary telescope. Methods of photographic registration by means of a motion picture camera were tried. Various sources of light were employed, including sunlight and the electric arc. Finally an arrangement was perfected for making observations with an astronomical telescope having an objective of five inches aperture and a magnification of fifty diameters. The source of light adopted was a large acetylene lamp of the kind commonly used for automobile headlights. An extended series of experiments was made to determine the influence of inequality of temperature and of radiant heat, and various insulating covers were provided for the base of the interferometer, and for the light path. These experiments proved that under the conditions of actual observation the periodic displacement could not possibly be produced by temperature effects. An extended investigation in the laboratory demonstrated that the full-period effect mentioned in the preliminary report on the Mount Wilson observations is a necessary geometrical result of the adjustment of mirrors when fringes of finite width are used and that the effect vanishes only for fringes of infinite width, as is presumed in the simple theory of the experiment.
In July, 1924, the interferometer was taken again to Mount Wilson and mounted on a new site where the temperature conditions were more favorable than
those of 1921. The interferometer house was also mounted with a different orientation. Again the observations showed a definite positive effect corresponding to the observations previously made at Mount Wilson. The observations on Mount Wilson were resumed in March, 1925, and continued until about the middle of April, during which time 1,600 measures of the drift in 35 sets were made. Again many variations in detail of arrangement of parts and in methods of observing were made without in any way altering the result. Throughout the latter epoch of observations, the conditions were exceptionally good. Some of the time there was a fog which rendered the temperature very uniform. Four precision thermometers were hung on the outside of the house. On several occasions the extreme variation of temperature was not more than 0.1 degree and usually it was less than 0.4 degree. Such variations did not at all affect the periodic displacement of the interference fringes. The observations of April, 1925, give results almost identical with those of April, 1921, notwithstanding that the interferometer had been rebuilt and that a different system of illumination and different methods of observation were employed and that it was mounted on a new site in a house differently oriented.
The interferometer readings, being plotted, give directly by harmonic analysis (carried out with the mechanical harmonic analyzer) the azimuth and magnitude of the ether drift. There are no corrections of any kind to be applied to the observed values. In the work so far, every reading of the drift made at Mount Wilson has been included at its full value. No observation has been omitted because it seemed to be poor, and no “weights” have been applied to reduce the influence on the result, since no assumption has been made as to the expected result. It may be added that while the readings are being made, neither the observer nor the recorder can form the slightest idea as to whether any periodicity is present, much less as to the direction or amount of such periodicity.
The test of these observations is whether they lead to a rational and wholly consistent indication of a constant motion of the solar system in space, combined with the orbital motion of the earth and the daily rotation on its axis. There is a specific relation for a given latitude between the observed azimuth of drift and the sidereal time of observations. Observations at different sidereal times should show different azimuths and all observations at the same sidereal time should show the same azimuth for a given epoch. A preliminary graphical solution of the observations indicates that these conditions are fulfilled.
It need hardly be said that the determination of the absolute motion of the solar system from such interferometer observations is one of great complex