time

And this was the subject of separate discussion at the conference on MMX in Pasadena, 1927; it seems impossible Roberts to be unaware of it, the more that before he wrote this paper, one of the authors cited him the shorthand reports of this conference, drawing his attention to the half-turn and full-turn effects and to the ways how Michelson, Miller and other experimenters processed their data [71]:

“Prof. Lorentz: … Though Miller states that he succeeded to exclude this effect to a great extent in his measurements in Cleveland, and it is easily explainable in the experiment, I would like to grasp the reasons clearer. Speaking at the moment as a supporter of theory of relativity, I have to state this effect to be not existent at all. Actually, the turn of device as the whole, including the source of light, gives no shift from the view of relativity theory. There has to be no effect when the Earth and device rest. After Einstein, the same absence of effect has to be observed for moving Earth. The effect of full turn, thus, is inconsistent with the theory of relativity, and it is of great importance. If then Miller revealed the systematic effects whose existence we must not negate, it is important to get to know the cause of full-turn effects” [72, p. 161].

“Miller: The Hicks theory bears in mind the fact that practically the image c (Fig. 6.18) of the mirror a accounts that a is slightly inclined to b. This is absolutely necessary to obtain the rectilinear beams of fringes of finite width. Thus, his criticism is inapplicable to the real case. When b and c are inclined to each other, the real ethereal wind will give an additional effect predicted by Hicks that is periodical in the full turn of device. Hicks has calculated the fringe shift, showing it to be independent of the angle between b and c. The effect grows with the growing angle and narrowing fringes. If the sought shift due to the ethereal wind has to be periodical in each half-turn, then we correctly excluded the full-period shift of fringes. This was done by graphical representation of single observations, turn by turn of interferometer; these curves have been analysed by mechanical harmonic analyser and the second harmonic (the half-turn effect) was presented as the consequence of ethereal wind. In presence of ethereal wind affection, there necessarily is present the effect of full turn after Hicks, and its presence can be thought one more evidence of presence of ethereal wind. The value and phase of effect of full-period shifts vary, as they depend on adjustment of mirrors equally with the ethereal wind. There were shown slides that represented the full-period effect. Obviously, the shift of fringes is different for different runs of observations. On the other hand, the half-period effect is characterised by the constant value. The full-period shift is small, while the width of fringes is such that five of them cover the mirror having 10 cm in diameter. At other conditions, however, the shift can be very large. The full-period effect is not new, it was present always in all experiments. It is also present in the Michelson’s primary experiments” [72, p. 168–169]. And the reflection of this systematic that Michelson took as an experimental error we really find already in his paper of 1881: “The first number is too small to consider it as a typical for the shift produced by a simple change of direction, and this last has to be zero. These numbers are simply the errors of experiment. Actually, it has to be seen from the final numbers of columns, as the numbers increase (or decrease) more or less regularly from left to right. This gradual change in no case has to affect the periodic change that we seek and that itself will exclude this error, simply because the sum of numbers of two columns from the left has to be less (or more) than the sum of numbers in the right columns. This is enough corroborated by the fact that there where the excess is positive (negative) for the columns N and S, it is positive (negative) also for N – E, S – W. Consequently, if we can exclude the monotonous change, we can expect a considerable decrease of the error” [73, p. 15].

Following the data of this Michelson’s paper, nothing to speak of trustworthiness of systematic shift, as in the given tables there is absent the last column related to the north orientation of device without which, after the intermediate data, it is hard to judge. But after the data of Michelson’s paper of 1887 [67], we can follow more specifically the pattern of systematic shift of fringes, as the table of this paper, of six runs made at different time, contains the data of full turn of the interferometer. We will not process these data but will only group them by dates of measurements. As the first three runs relate to the diurnal measurements of 8, 9 and 11 July, and the second three runs – to the night measurements on the same days of year, let us join the diurnal and night measurements of same day, see Fig. 6.19.

In these graphs we see that all runs of measurements have the systematic shift of fringes. We also see that the systematic shift related to the full-turn effect could be equally increasing either decreasing, which corroborates Miller’s conclusion that the systematic shift depends on the way, how the device is adjusted. The fact that in 1887 Michelson and Morley gave not a great importance to the full-turn effect, thinking it to be the error of device, allowed them to adjust the device in different ways, which naturally reflected on the pattern of change of systematic shift. But the fact that such shift is not the systematic error related to the disadjustment of the device was corroborated by the positive results of experiment on the stationary interferometer that Michelson carried out with Gale in 1925: “The shift of fringes on account of Earth’s rotation has been measured by different observers at different days with a complete readjustment of mirrors; the reflected image was sometimes to the right, sometimes to the left of transmitted image (which corroborates the possibility of both increase and decrease of the systematic shift, dependently on the conditions of adjustment – Authors). The deviation usually was averaged … Noting the difficulty of observation, we have to establish that the observed and calculated shifts agree in limits of observational error” [74, p. 60–61].

It would be useful to show here the affection of masking effect onto the interference techniques to reveal the aethereal wind. The matter is, before the experiment of 1925, Michelson made a similar experiment described in his paper [75], see the scheme of this experiment in Fig. 6.20.

“The light of the source S of the calcium light or electric arc lamp has been divided into two narrow beams on the glass, lightly silvered plane-parallel plate o. Two beams were reflected by two mirrors along the paths oabcoe and ocbaoe relatively. Two paths are equal, the fringes can be observed with the telescope at e” [75, p. 475].

As we can see, this scheme is fully symmetrical and any deflections of beams on the sides of rectangle due to the motion of device through the aether as the whole will find the similar shifts of the meeting beam in the symmetrical arms. So we have no grounds to expect some clear effect. So this experiment necessarily had a negative result which Michelson yielded: “The conclusion of these results is that if there is some shift of fringes, it is less than one seventh of the width of fringe e” [75, p. 477]. Though some side effect related first of all to the necessity to produce a little angle between the beams to form the interference pattern had to be present, of course, and Michelson attributed it to the difficulties of experiment: “We encountered the difficulty when choose the starting mark. The double image of the source does not remain on the transverse hairs of observing telescope for any large interval of time, irrespectively of caution of double reflections at the angles. But the use of this double image itself as the starting mark excludes any possible errors because of diurnal variations of temperature etc. This double image and fringes are not in the focus simultaneously, but through a very little sacrifice in determining of each, the measurements can be made with an essential accuracy” [75, p. 477]. As it follows from the citation, the effect was that they could not achieve a coincidence of the source and fringes at a time, and that the image was double. Naturally, even if Michelson tried to study this effect, due to the dependence of the value of effect and even of the direction of shift of the pattern on the adjustment of interferometer, Michelson would yield the statistically spread data similar to those which all researchers yielded, including himself, in the trivial (rotating) interference scheme with the full-turn effect. And though the effect in the considered first Michelson’s scheme is not directly the full-turn effect in the rotating interferometer, the meaning of this effect is same, as in both cases it is caused by the necessity to introduce little angles of disadjusting of mirrors to yield the interference pattern, due to which the affection of aethereal wind in both cases was so much dependent on the adjustment of device in each run of the experiment.

In his second version of stationary interferometer, Michelson essentially changed the scheme, see Fig. 6.21.

As we see, this scheme has already two loops. “The lines we had to measure were produced by the beams that passed in opposite directions about ADEF .. As the starting marker from which we had to measure we formed the second row of lines with help of system of mirrors ABCD. To find the shift of lines, ABCD had a well less perimeter, and recorded shifts were the actual shifts between the central lines of two rows. Generally, two rows of fringes will not coincide in their location, absolutely independently of whatever shift of ether or Earth’s rotation, until two direct and two reflected images will not be fully superimposed. The central fringes of the row formed by the mirror of short loop will be in the middle between the direct and reflected image of the source, and the central fringe of the long loop has to be in the middle between the direct and reflected images, if the affection of Earth’s rotation is absent” [74, p. 59].

This asymmetrical scheme, as we cited above, gave a positive result, quite expected just due to asymmetry. Actually, let us draw attention, Michelson converged the coupled images of each loop formed by the opposite beams of each loop. In this way he made each loop insensitive to the turn of interferometer relative to the aethereal wind. With it he got rid of effect that arises due to the presence of the angle of interference between the beams in the previous scheme. The measurements were conducted with account of difference in distortions that arise in the loops of different size when the Earth turns relative to the aethereal wind. Naturally, in this device Michelson was able to measure only the motion of the Earth’s surface due to its rotation, as the data yielded by Michelson registered only the difference of speeds with and against the aether’s motion. In such subtraction, the Earth’s absolute motion was compensated. But the very fact of affection of this motion related to the revolution in the aether fully corroborates also general motion of the Earth through the aether. And this experiment shows that even if there exists the deceleration of aether near the Earth’s surface either its screening by metal, concrete etc., it is insignificant. Michelson carried out his experiments in the vacuum chamber whose screening from the aethereal wind is thought to be considerable, and the average statistical results of measurements differ from each other only by 2,54 %. While in case of screening or presence of boundary layer, the results would differ in times, as we could see it above, considering the experiments by Miller and by Michelson with the turning interferometer. It follows from this that a large statistical spread and little value of results in the rotating scheme are caused by an effective masking effect caused, in its turn, by the arrangement of the very scheme of rotating interferometer. On these grounds some time ago we suggested to Dr Yury Galayev to build up the experiment just on the asymmetrical scheme. In Fig. 6.22a we show the suggested scheme, and in Fig. 6.22b – the scheme which he realised in the radio range.

Comparing the schemes, we see, they both are based on the same idea. The only difference is, the scheme in Fig. 6.22 suggested to measure in the range of visible waves, and Yu. Galayev has experimented in the radio range, which gave some difference. But just asymmetrical scheme gave a positive effect.

We see from analysis that the masking effect is so much effective in the interferometric scheme that the negative result is factually predestined in symmetrical schemes. So the affection of aethereal wind can reveal itself only as a parasitic effect related to the disadjustment of mirrors that makes an angle between the beams. But in asymmetric schemes the affection of device’s motion relative to aether is robustly registered. We can even suppose: should in the early 20th century there existed stable coherent sources of light with which they could make experiments on Michelson’s interferometer with much different lengths of arms, the positive result would be detected already then.

Unfortunately, in the early past century such sources did not exist; available sources disabled the researchers to make much different length of arms, and on equal arms the effect revealed itself just in the full-turn effect related to the change of Earth’s orientation during the turn of interferometer. With it, as we see, all interferometric experiments with a moving device showed its presence. Different researchers avoided this effects in different ways. Michelson subtracted it immediately as a parasitic effect. Illingworth readjusted his device after each run. Only Miller experimentally and Hedrick theoretically followed the systematic pattern of fringes shift and showed, this effect is not an occasional error but the phenomenon related to the turn of interferometer in space during the full turn of device. And this systematic shift is well seen in Fig. 6.17.

With account of this systematic, the full-turn effect can be explained in no other way as the displacement of device relative to the aethereal wind during the full turn. In that number, this phenomenon cannot be substantiated as the revelation of Sagnac effect. For it, it would be necessary, the interferometer to rotate with the speed well more than those 30–40 seconds (and in first Michelson’s experiments even 6 minutes!) per one turn that were spent in practical study.

The more, this effect cannot be explained within Einsteinian conception, – and Lorentz was right when said it at the Pasadena conference. Actually, if, according to Einstein, the revolving frame is equivalent to the inertial frame and in each such frame the lightspeed is isotropic, any turns of interference device in which the device turns to its initial location are unable to cause the systematic shift of interference pattern. So the theory of relativity can predict only zero result. And if the adherents of Einsteinian conception, analysing the existing experimental data, agree that this systematic is present in the experiments on revelation of aethereal wind, so they, irrespectively of what they think about it, admit the positive result of interference techniques of finding the aethereal wind.

As we already said, a half-turn effect is masked by the opposite shift of beams in the device which compensates the reciprocal change of paths that beams pass: “… the turn of device by 90^? in an ideal experiment causes no effect, as, despite the exchange of distances which two beams pass, their positions exchange at the same time; thus, the beam with a longer path has the same position relatively the beam with a shorter trajectory after rotation as before. Thereupon, the pattern of fringes after a turn is indistinguishable of that which was before the turn” [72, p. 149]. This is the cause of a little amplitude of the effect, because of which relativists are sure of negative result. And now this is already the problem of interferometric techniques to seek the aethereal wind. The difficulty arises because when turning the device, the beams do not remain at the trajectories which Michelson and then Lorentz predicted. They had to incline the direction of source’s in order, in the moving frame the beams to propagate normally to each other. When the device turned, this necessity will cause the fact that the beams in the interferometer already will not be perpendicular to each other, as the angle of transverse beam inclination already will not be compensated by the motion of the device as the whole. At the same time the longitudinal beam will not have a necessary inclination, to be perpendicular to the previous transverse beam in the moving frame, as we showed it in [14, p. 53]. But the fact is obvious: the experiments revealing the aethereal wind were positive from the very beginning and the results of these experiments fully debunk the postulates of theory of relativity. In that number they debunk as invalid the relativistic identification of revolving inertial frame with that inertial, which additionally corroborates the results of our study in this subsection.

SELF	72
S.B. Karavashkin and O.N. Karavashkina