D O C U M E N T 5 4 S E P T E M B E R 1 9 2 7 6 5 is also invariant, since transforms as The left-hand side of the equation behaves as follows: the equation[3] is an invariant equation owing to the invariant nature of this equation, and in fact this holds for all arbitrary canonical transformations. Thus, also the equation . [is invariant]. The left side of this equation transforms as , thus also as . Therefore, it holds that is an invariant with respect to arbitrary transformations. It therefore follows that the equation[4] is invariant under all transformations that leave invariant. We can hardly ask for more. It is thus quite justifiable that we should make an attempt at the problem of the rotator on the basis of this equation. With warm greetings, yours truly, A. Einstein 54. To Chaim Herman Müntz [Berlin,] 17 September 1927 Dear Dr. Müntz, Many thanks for your message. Unfortunately, I cannot accept that your conclu- sions are correct. That the excess radius that you calculate, H p i dq i ------- dq i dt q i ------------------- - -------i dp dt p i ------------------- - + t = q i H p i p i H q i – t = t ------ 1 -- q i -------------- H p i p i --------------H q i – q i -------------- H p i p i -------------- H q i – j------p 2 h i H p i ------- = p i dq i