D O C U M E N T 4 2 0 N O V E M B E R 1 9 2 6 3 9 7 , where α is substituted by . For small s’s, one can set Hence If is very small also compared to , which one needs in order to arrive at large n’s, then n0 becomes very large compared to the other ns’s, which ¢really is the only way, too,² simply amounts to the same thing as the “condensation.” The equations of motion for singularities can really be derived relativistically. But it appears as if nothing at all “unclassical” results from it.[4] — Now to something else that I tell you only with much regret. Because of the League of Nations and several industrial matters that I have gotten myself into, I have so little energy & time left that I cannot maintain my position in Leyden. I ear- nestly ask you please to tell me the right form in which I can sever this tie ¢in a friendly manner². I have also already become too unproductive to allow this ar- rangement to appear justified. So don’t scold me but help me finish up this matter in a nice and friendly way.— Our experiment on the Compton effect still isn’t done, but will surely succeed.[5] In the meantime, I recommend that you clear other channels besides the ones clos- est to your purview. Yours, Einstein Manuscript by George Uhlenbeck and Paul Ehrenfest Manuscript f[or] J[ournal of] Wretched Physics[1] Does the Bose-Einstein statistics for ideal gases lead to a degenerative conden- sation? G. E. Uhlenbeck and P. Ehrenfest. (Leyden) Motto: When a sovereign builds, canal cleaners also get work. ns 1 αeβεs 1– --------------------- -= 1 α′ + αeβεs 1 α′)( + ( 1 βεs) + 1 α′ βεs + + = = ns 1 α′ βεs + ------------------- - for small s = n0 1 α′ ----- = n1 1 α′ βε1 + -------------------- = etc. n0 n1 » α′ βεs