April 30, 1926]
SCIENCE
437
tended investigation in the laboratory demonstrated that the full-period effect mentioned in the preliminary report of the Mount Wilson observations is a necessary geometrical consequence 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 real periodic displacement of the fringes, as in all the observations previously made at Mount Wilson and at Cleveland.
In spite of long-continued efforts it was impossible to account for these effects as being due to terrestrial causes or to experimental errors. Very extended calculations were made in the effort to reconcile the observed effects with the accepted theories of the ether and of the presumed motions of the earth in space. The observations were repeated at certain epochs to tests one after another of the hypotheses which were suggested. At the end of the year 1924, when a solution seemed impossible, a complete calculation of the then expected effects, for each month of the year, was made for the first time. This indicated that the effect should be a maximum about April first, and further that the direction of the effect should, in the course of the twenty-four hours of the day, rotate completely around the horizon. Observations were made for verifying these predictions in March and April, 1925. The effect was equal in magnitude to the largest so far observed; but it did not point successively to all points of the compass, that is, it did not point in directions 90° apart at intervals of six hours, nor point in opposite directions at intervals of twelve hours. Instead of this, the direction merely oscillated back and forth through an angle of about 60°, having, in general, a northwesterly direction.
Previous to 1925, the Michelson-Morley experiment has always been applied to test a specific hypothesis. The only theory of the ether which has been put to the test is that of the absolutely stationary ether through which the earth moves without in any way disturbing it. To this hypothesis the experiment gave a negative answer. The experiment was applied to test the question only in connection with specific assumed motions of the earth, namely, the axial and orbital motions combined with a constant motion of the solar system towards the constellation Hercules with the velocity of about nineteen kilometers per second. The results of the experiment did not agree with these presumed motions. The experiment was
applied to test the Lorentz-FitzGerald hypothesis that the dimensions of bodies are changed by their motions through the ether; it was applied to test the effects of magneto-striction, of radiant heat and of gravitational deformation of the frame of the interferometer. Throughout all these observations, extending over a period of years, while the answers to the various questions have been “no,” there has persisted a constant and consistent small effect which has not been explained.
The ether-drift interferometer is an instrument which is generally admitted to be suitable for determining the relative motion of the earth and the ether, that is, it is capable of indicating the direction and the magnitude of the absolute motion of the earth and the solar system in space. If observations were made for the determination of such an absolute motion, what would be the result, independent of any “expected” result? For the purpose of answering this general question, it was decided to make more extended observations at other epochs in 1925, and this was done in the months of July, August and September.
It may be asked: why was not such a procedure adopted before? The answer is, in part, that we were concerned with the verification of certain predictions of the so-called classical theories; and in part that it is not easy to develop a new hypothesis, however simple, in the absence of direct indication. Probably a considerable reason for the failure is the great difficulty involved in making the observations at all times of day at any one epoch. I think I am not egotistical, but am merely stating a fact when it is remarked that the ether-drift observations are the most trying and fatiguing, as regards physical, mental and nervous strain, of any scientific work with which I am acquainted. The mere adjustment of an interferometer for white-light fringes and the keeping of it in adjustment, when the light path is 214 feet, made up of sixteen different parts, and when it is in effect in the open air, requires patience as well as a steady “nerve” and a steady hand. Professor Morley once said, “Patience is a possession without which no one is likely to begin observation of this kind.” The observations must be made in the dark; in the daytime, the interferometer house is darkened with black paper shades; the observations must be made in a temperature which is exactly that of the out-of-door air; the observer has to walk around a circle about twenty feet in diameter, keeping his eye at the moving eyepiece of the telescope attached to the interferometer which is floating on mercury and is turning on its axis steadily, at the rate of about one turn a minute; the observer must not touch the interferometer in any way, and yet he must never lose sight of the interference fringes, which are seen only through the small
438
SCIENCE
[Vol. LXIH, No. 1635
aperture of the eyepiece of the telescope, about a quarter of an inch in diameter; the observer makes sixteen readings of the position of the interference fringes in each turn, at times indicated by an electrical clicker; these operations must be continued without a break through a set of observations, which usually lasts for about fifteen or twenty minutes, and this is repeated continuously during the several hours of the working period.
When observations are in progress the interferometer to which the observing telescope is attached is caused to rotate on the mercury float so that the telescope points successively to all points of the compass, that is, it points to all azimuths. A relative motion of the earth and the ether should cause a periodic displacement of the interference fringes, the fringes moving first to one side and then to the other as referred to a fiducial point in the field of view, with two complete periods in each rotation of the instrument. Beginning when the telescope points north, the position of the fringes is noted at sixteen equidistant points around the horizon. The azimuth of the line of sight when the displacement is a maximum having been noted at two different times of day, it is a simple operation to calculate the right ascension and declination, or the “apex” of the presumed “absolute” motion of the earth in space. The determination of the direction of the earth’s motion is dependent only upon the direction in which the telescope points when the observed displacement of the fringes is a maximum; it is in no way dependent upon the amount of this displacement nor upon the adjustment of the fringes to any particular zero position. As the readings are taken at intervals of about three seconds, the position of the maximum is dependent upon observations covering an interval of less than ten seconds. A whole period of the displacement extends over only about twenty-five seconds. Thus the observations for the direction of the absolute motion are largely independent of ordinary temperature disturbances. The observation is a differential one and can be made with considerable certainty under all conditions. A set of readings usually consists of twenty turns of the interferometer made in about fifteen minutes’ time; this gives forty determinations of the periodic effect. The forty values are simply averaged to give one “observation.” Any temperature effect, or other disturbing cause, which is not regularly periodic in each twenty seconds over an interval of fifteen minutes would largely be cancelled out in the process of averaging. The periodic effect remaining in the final average must be real.
The position of the fringe system is noted in units of a tenth of a fringe width. The actual velocity of the earth’s motion is determined by the amplitude of the periodic displacement, which is proportional
to the square of the relative velocity of the earth and the ether and to the length of the light path in the interferometer. A relative motion of thirty kilometers per second, equal to the velocity of the earth in its orbit, would produce a displacement of the fringes from one extreme to the other, of 1.1 fringes. Disturbances due to temperature or other causes lasting for a few seconds or for a few minutes might affect the actual amount of the observed displacement and thus give less certain values for the velocity of relative motion, while at the same time the position of maximum displacement is not disturbed. Thus it is to be expected that the observations for the velocity of motion will not be as precise as the observations for the direction of motion. The two things, magnitude and azimuth of observed relative motion, are quite independent of each other.
It is desirable to have observations equally distributed over the twenty-four hours of the day; since one set requires about fifteen minutes of time, ninety-six sets, properly distributed, will suffice. The making of such a series usually occupies a period of ten days. The observations are finally reduced to one group and the mean date is considered the date of the epoch. The observations made at Mount ^Wilson in 1925 correspond to the three epochs, April 1, August 1 and September 15, and are more than twice as numerous as all • the other ether-drift observations made since 1881. The total number of observations made at Cleveland represent about 1,000 turns of the interferometer, while all the observations made at Mount Wilson previous to 1925 correspond to 1,200 turns. The 1925 observations consist of 4,400 turns of the interferometer, in which over 100,000 readings were made. A group of eight readings gives a value for the magnitude and direction of the ether-drift function, so that 12,500 single measures of the drift were obtained. This required that the observer should walk, in the dark, in a small circle, for a total distance of 100 miles, while making the readings. Throughout these observations the conditions were exceptionally good. At times there was a fog which rendered the temperature very uniform. Four precision thermometers were hung on the outside walls of the house; often the extreme variation of temperature was not more than one tenth of a degree, and usually it was less than four tenths of a degree. Such variations did not at all affect the periodic displacement of the fringes. It may be added that while the readings are being taken, neither the observer nor the recorder can form the .slightest opinion as to whether any periodicity is present, much less as to the amount or direction of any periodic effect.
The hundred thousand readings are added in groups of twenty, are averaged and then are plotted in curves. These curves are subjected to mechanical harmonic