2 7 6 D O C U M E N T 2 6 2 A P R I L 1 9 2 6 The wave theory de- mands that clear in- terferences appear for a difference in path length l in the case of the compensating tilt of the mirrors [6] lens with short focal length f [7] telescope set at [8] interference apparatus Already for moderate l this would be of appreciable magnitude for small focal lengths. b) The grating is installed directly behind the window. [9] possible, but not necessary [10] interf. app. with parallel plates Then interference apparatus either without lens or (if one wants to select part of the light source) with a setup of lenses that corresponds to a telescope set at ∞. If a) gives a positive result, one can also use the setup a) with tilted mirror compensation.[11] Kind regards, your A. Einstein P.S. In order to prevent the diffraction at the grating from destroying the success of b, the grid must not be positioned farther away than 1 cm from the canal ray. If this is not feasible, one will again have to form an image of the canal ray on the grating. [6] [7] [8] ∞ 1vl 2cf -- - -- - [9] [10]