5 4 D O C U M E N T 3 0 J A N U A R Y 1 9 2 2 The ray thus moves in a circle. As [emerges] from our earlier considerations The radiation emitted at time therefore propagates in a bent line, to be pre- cise, on a circular path. The wavelength normal that according to one of our earlier results has the direction of the radius vector drawn to the starting point hence undergoes a deflection in the sense of a growing x of magnitude . . . . (14) This result is equivalent to our equation (3). To see this, one only has to take into account that the beam’s rotational velocity is outside of the dispersive dielectric inside it, therefore, it is equal to , which term is set equal to . Furthermore, for one has to set length l of the dispersive layer and introduce for V the refrac- tive index n, for the frequency ν. Thus rigorous proof is provided that equation (3), which is not confirmed by experience, is in fact a consequence of the undulatory theory. 30. From Paul Ehrenfest [Leyden,] 19 January 1922 Dear Einstein, To facilitate reading this letter: Borrowing a remark by Gibbs (1886), I believe I can show: even if one grants you the rotation of the wave planes as proportional to the length of the bisulf[ide]- of-carb[on]-tube, there follows (on a purely classical basis) a negative result for your experiment on a classical foundation.– The point is: group of waves![1] Sheets A, B, C reiterate more clearly what pages 1, 2 present somewhat messily. Forgive me if I am mistaken. Please reply to me, if only briefly and provision- ally. Don’t forget, either, to indicate the address of the Spanish professor, about whom I wrote yesterday on my postcard.[2] Don’t be angry with me if I am wrong don’t be angry with me if I am right. With warm regards, yours, P. Ehrenfest t 0= r0 r0ndω 1 -------- - dn - γ⋅ q Δ --- q Δn ------ - γ--- V ω - r0 dω ω