1 8 2 D O C U M E N T 1 6 2 J A N U A R Y 1 9 2 6 162. From Walter Schottky[1] Rostock, Physics Institute, 12 January 1926 Dear Mr. Einstein, I would like to interrogate you about one point in your paper, Z[eit]s[chrift für] Phys[ik], 20, [p.] 301, on the quantum theory of radiative equilibrium.[2] There an “inverse process” is supposed to be understood as a process in which a quantum is emitted in the same direction as the direction from which it is absorbed in the corresponding process.[3] For the relevant probability coefficients, (1) should hold. Now, I cannot make this definition of the inverse process rhyme with a classical analogy of the law (1) that I found, of the following kind. If b refers to instead of to ρν [ρν ], is the number of incident quanta of the cor- responding kind, then (1) becomes: (2) This relation has the same form as one of the classical laws derived from Rayleigh’s reciprocity laws,[4] which I call a “law of low reception,” which, however, states that the (most variously defined) absorption capa- city for radiation coming from a particular direction stands in this relation to the emission capacity in the relevant direction, with a preference for [the] low frequencies. In cases where the structure, whose determining radiation environment can be arbitrarily large, essentially only radiates in one direction (such as, for higher fre- quencies in the case of a loudspeaker with a long horn), it is very clear that a rela- tion of form (2) can only be valid for opposite directions of the incident and emitted radiation.[5] Shouldn’t the same have to be assumed in quantum theory as well? Re- versibility of unidirectional revolutions in the radiating system would not, to my knowledge, need to be inferred such an unidirectionality would just mean that the structure possesses preferred absorption and emission for waves of a specific cir- cular polarization. As I have pretty much finished the remaining chapters of my low-level reception work, of which I hope a part can be printed in the Academy’s reports,[6] I would be grateful if you would send me a brief note about this quantum question. With kind regards, yours sincerely, Walter Schottky a b -- - 8πhν3 c3 --------------- = Kν hν ------ 8π c ------ Kν = Kν hν ------ b a -- - λ2 =