2 9 2 D O C . 3 4 8 M A R C H 1 9 2 0 May I also note on this occasion that although I emphatically contradicted (in 4 or 5 public debates) local astronomers (with their blind and often uncritical declarations about your theory), I do very much admire your theory of gen. rel. and gravitation and regard it as a genuine analytical masterpiece. [Even quite irrespec- tive of whether or not it agrees with experience in all respects.][3] (It is perhaps bold on my part to take the liberty of passing judgment on your great work, even though I have worked very hard since 1912 to introduce, defend, and propagate precisely your spec. theory of rel. here in England.)[4] Although I dug the “ether”—which to me personally seems very improbable—out of 20 years’ worth of dust, I devote much more time & effort to various problems treated from the point of view of your relat.-gravit. theory,—and such problems have a special attraction to me. Some of it is now being published. I did suggest one problem (which actually comes down to filling an important gap in your theory and which I tackled in vain) to Dr. Kottler (Vienna) a few months ago[5] he wrote me recently, however, that he has not been able to deal with it until now. It regards, briefly put, the incorporation of the (suitably modified) lever prin- ciple (more generally, statics and kinetically coupled masses ≈ weights) into your theory. How would you approach such a problem? [What new principle should be introduced to treat it?—for in your theory such does not seem to exist yet.][6] Let’s assume two point-masses are “rigid” (cf. infra[7] ) (massless connecting rod). One point on this rod is fixed. What is the criterion of equilibrium for this system sus- pended in a given gravitational field? More generally, if no point is fixed, how does the system move (“fall”)? [It would suf- fice for me to consider the case in which the contribution of to were negli- gible.][8] Evidently your principle does not suffice.[9] How can one extend it for the system ? (More generally, for any “rigid” system of mass-points?) I must add how I understand “rigid.” I would suggest the following definition, which I have generalized from Herglotz’s special relativistic definition of rigidity.[10] Let there be in the given field F , and let be conceived as world lines of (initially conceived as massless points). Now, if for all possible motions of the pair of points, are equidistant curves [“distance” mea- sured from (1)],[11] I say that is rigidly connected with Assuming this definition, how does the rigid system (now) move for masses equipped with points (in “natural measure,” say)? How should the principle m1, m2 m2 m1 m1, m2 m1, m2 gικ δ∫ sd 0= m1m2 1)ds2 ( gικdxιdxκ = L1, L2 m1, m2 L1, L2, m1 m2.— m1, m2