The Kennedy-Thorndike experiment

The only difference between Kennedy-Thorndike's interferometer and Michelson's interferometer is the following: lengths of perpendicular arms were made evidently different in Kennedy-Thorndike's interferometer. However, for the interferometric pattern it is only important the difference of paths of light rays with respect to the wavelength of used light (part of the wavelength). Besides, an interferometer arms (for example, Michelson's interferometer) are always measured with an accuracy which is less, than the wavelength of used light. Therefore, contrary to the judgement of [38], the Kennedy-Thorndike experiment does not principally differ in anything from the Michelson-Morley experiment. As a result, all remarks to the Michelson-Morley experiment, which are indicated previously, will be remained common for both these experiments.

If one proceeds from the experiment tasks (on detecting the effect of the interferometer system motion on the speed of light), then author's estimate of $v\le 15$ km/s is more adequate, than that stated in the textbooks, though it is incorrect too (see below). The great stability in temperature, beginning with some limit, does not matter, because at any $T=constant$ ($T\ne 0$) always exist temperature fluctuations and oscillations of a crystal lattice of the base. Of most importance is the fact, that various speeds of light $c(\omega)$ (the only possible distinction - see above) have not been compared for various frequencies $\omega$, which would be impossible to be done in a similar experiment. Besides, for the empty space all classical considerations for inertial systems remain valid; that is, Galileo's principle of relativity [48] is met in this case. The general notion about metal screening for the ether model is applicable to this experiment as well. Thus, all listed experiments have no relation even to detecting the motion of the Earth.