D O C U M E N T 2 7 7 M A Y 1 9 2 6 2 9 1 277. To Emil Rupp [Berlin,] 5 May 1926 Dear Mr. Rupp, I like the outline of the experimental setup you proposed on 1 May [1] however, it seems to me that you could do without the lenses and position the interference setup directly behind the grating. The interference has to reappear in all path dif- ferences when rotating the grid by 90°, so that it is parallel to the canal ray. The other experiment[2] would have to be con- ducted in a manner where the grating is replaced by a lens. In that case, optimal interference should be achieved by turning the mirror. You can turn ei- ther one of the mirrors S1, S2, or mirror S0 , but the latter only by half the amount, which is a disad- vantage for the measurement. In this experiment you should use interferences of equal inclination. Theoretically, these are much more transparent. No narrow slits are necessary in the experiment with the lens. The great thing about it is that the Doppler components are compensated by the position of the mir- rors. However, I am almost certain that the expected results will conform to the wave theory. The expectation expressed in my note can hardly apply.[3] I am sure you will agree when you see my simple and convincing theoretical thoughts. Kind regards, your A. Einstein Translators’ note: Based on a translation by Doris Lonk and Tilman Sauer.