2 1 0 D O C U M E N T 1 9 4 F E B R U A R Y 1 9 2 6 194. To Paul Ehrenfest [Berlin,] 12 February 1926 Dear Ehrenfest, First of all, it seems to me that Holst should be chosen as first in line.[1] His achievements in physics in Holland are unique, and he has sacrificed his scientific personality, so to speak, for the sake of extrapersonal interests. He is entitled to anything equivalent to what the others are given. If an additional person has to be chosen, I would not choose Ornstein,[2] because one must consider not only personal merit but also the Academy’s interest you know that I greatly value moral issues. The influence of a man who is too easily guided by selfish motives should not be expanded. As regards Coster and Cramers,[3] I share your opinion. Scientifically, they are worth more than Fokker[4] (in my opinion), although I’m hesitant about Tetrode.[5] Zernike[6] is capable but, as far as I know, has actually never had a properly independent idea of his own. I haven’t been struck by any great personal qualities in him, either. Therefore I defi- nitely would not select him. Fokker and Tetrode are left over. Fokker is a very clear thinker and a valuable human being but certainly less inventive than Tetrode. Besides, one should not give up any attempt to draw Tetrode out of his hiding place a bit. There definitely is a genius streak in him, even though he has only produced one thing of lasting value.[7] That exaggerated solitude, which is surely pathological, is certainly dam- aging his creativity. I would choose Tetrode because the most good would be ac- complished. Fokker should be considered at the next opportunity such recognition for him would certainly be in order.— I have been greatly occupied with the Heisenberg-Born idea.[8] Despite much admiration for this idea, I am tending more and more toward considering it inappro- priate. There can’t be any zero-point energy in cavity radiation. I deem the pertinent argument by Heisenberg, Born, and Jordan (fluctuations)[9] as faulty, if only be- cause the probability of large fluctuations (for example, arrival of the total energy in partial volume V from ) certainly does not come out correctly that way. Fur- thermore, the nonvanishing of the angular momentum does not agree with the quan- tum rules for systems (pμqν ). Finally,—it seems to me—the theory for one degree of freedom isn’t even covariant with coordinate transformations. If is a Hamiltonian function of a pointlike motion, then I can introduce a new coordinate and the associated P. In ordinary mechanics a function belongs to these variables, such that . If this problem is solved for the new variables PQ with matrices, in the Heisenberg- V0 pνqμ – 0 = H p q) ( Q ϕ(q) = H* H p q) ( , H*(P, Q) =