D O C U M E N T 2 7 9 M A Y 1 9 2 6 2 9 9 gaps, then the path differences of the largest and smallest interference capabilities would be characterized by the equations , where n means a positive whole number (incl. 0). Result. If the statement about the lack of influence of a parallel shift of the light sources on interference phenomena of extended sources of light is correct, then the interference phenomena observable in homogeneous canal rays must proceed in accordance with the classical emission theory of light, i.e., as if the canal ray particles were Hertz oscillators. An influence of the quantum structure of the radi- ation would not be expected. Postscript. The present paper was written in May 1926 and served as a guide to Mr. Rupp for experiments described in the following paper.[15] They have fully confirmed the theory. 279. To Emil Rupp [Berlin,] 9 May 1926 Dear Mr. Rupp, I have now worked out my thoughts on the interference properties of the canal ray light. Enclosed you will find my paper,[1] which contains something new even for you, in regard to the setup where a lens is positioned in front of the interference apparatus. In my opinion, it would be best if my theoretical paper and your exper- imental one were published directly one after the other, for example in the Akademie-Berichten der preussischen Akademie or in the Zeitschrift für Physik. Do you agree? Kind regards, your A. Einstein 280. From Georg Joos Jena, Kaiser Wilhelmstr. 7a, 9 May1926 [Not selected for translation.] dmax 2n bc v ----- -, = dmin 2n 1+ ( ) bc- v ----- =