D O C U M E N T 7 1 S E P T E M B E R 1 9 2 5 7 9 corpuscule boundary anymore that was a misguided endeavor, I believe.[3] One will probably never arrive at a reasonable theory along the route of induction, even though I do think that experiments on the very principles, such as the Stern-Gerlach or Geiger-Bothe experiments,[4] can be of serious use. What you say about Bohr and me as opposed to other theoreticians with respect to your research doesn’t sur- prise me. There are mad defenders of principles and there are virtuosos. All three of us belong among the first sort, and (certainly the two of us, at least) have little virtuosity. Hence the effect when encountering utter virtuosos (Born or Debye):[5] discouragement. It works similarly the other way around, by the way. I haven’t received any request from the Astroph. Journ. yet. I would like to do Mrs. Julius a little favor.[6] What should I do? If you know some way, do send me the information and I shall do my utmost. What you convey to me as Lorentz’s reservations about Miller is smart and fine.[7] But 30° would be decisive in this subtle matter only if it could be demon- strated that the precision and inner consistency among the measurements sufficed. I also have an idea, albeit a very trivial one, about how Miller’s “bubble” could have formed.[8] Temperature differences in the air between the two beams, of an order of magnitude of °, would suffice to cause the whole fuss. There are surely systematic causes (solar irrad[iation] of the walls, heating, etc.) that are capable of producing such differences systematically. The refractive index is, of course, of or- der of magnitude . A fluctuation in the abs. temperature of ord. of mag. yields a relative difference of between the light paths. I didn’t see any remark about how such an error was avoided. They could be avoided by means of ventila- tors. I posed a question to Miller in this regard by letter. We shall see what he thinks of it. During the summer I wrote an article about gravitation/electricity (Academy’s Berichte) that is in itself very captivating, which I want to send you.[9] But now I very much doubt its truth again. I am doing a lot of calculating with Grommer,[10] without anything decisive pro or contra resulting from it so far. The balance is rather tilting toward the contra side, though. We must not, under any circumstances, let Burger be preferred over Kramers.[11] On whom does this depend? You and Lorentz can prevent that I shall naturally also immediately do my bit, provided it makes sense. Warm regards to all from your Einstein 1 10 ----- - 10 4– 10 3– 10 7–