2 1 6 D O C U M E N T 1 9 8 F E B R U A R Y 1 9 2 6 From this, one thus concludes that the quantity can be written as , where . One thus sees how subsequently the matrix character of x, y results, together with the frequency condition. The essential thing about this (formally not pretty) treatment of the rotator seems to me to be that one can spot exactly which physical hypotheses are plugged in during the quantization. What is done here for the rotator can be done for any other model, of course (oscillator, hydrogen), if cyclical variables are introduced. There- fore, the general result seems to me to be this: at first, quantization is not part of the formal apparatus of quantum mechanics, but the frequency condition certainly is. Two further hypotheses are at the basis of the quantization: 1.) There exists a normal state in which no radiation takes place. ¢(cf. my first paper, eq. p. 886).²[11] 2.) Radiation is absorbed and emitted only in amounts of hν. For the rotator, one would conclude that (1.) for the normal state , hence furthermore, that (2.) only the states (τ = a whole num- ber) really exist.[12] In the matrix formulation of quantum mechanics, hypothesis (2.) is put together with the combination relation, i.e., with the matrix theorem hypothesis (1.) always had to be assumed as a boundary condition in order to make the problem definite. I also think that it is prettier if these two hypotheses come subsequently as a con- sequence of the laws applying to the individual process. (For, it seems to me prob- able that quantum mechanics can never make direct statements about the individual process, but rather only yields mean values in the sense of the Bohr-Kramers-Slater theory.)[13] One can trace this very well in the collision processes I have been study- ing over the last few weeks. Pardon that my letter has become so long now I would be very glad if I could discuss sometime in person the problems broached here at the end Prof. Laue has invited me to speak at the colloquium in Berlin at the end of April [14] perhaps I may then also visit you once during those days. Yours sincerely, Werner Heisenberg td d eiϕτ) ( 2πi§ h -------- H p) ( H p τh·· 2π¹¹ ------ – © § – © eiϕτ. = eiϕτ e2πiνt ν H p) ( H p τh· 2π¹ ------ – © § – h -------------------------------------------- = ν 0= p 1 22π -------- - h = p τ 2¹ -- -+1· © § h 2π ------ =