D O C U M E N T 2 1 7 J U N E 1 9 2 8 2 1 5 terization of such a continuum, rather than the more complicated Riemannian cur- vature tensor. The simplest combinations that can be considered here are the vector and the invariants and .[18] From one of the latter (or from a linear combination of them), through multiplica- tion by the invariant volume element , (where h refers to the determinant |h | and dτ to the product dx 1 ... dx μ ), we can form an invariant integral J.[19] Setting yields 16 differential equations for the 16 quantities .[20] Whether or not this procedure leads to laws that are physically relevant will be the subject of later investigations.— It is instructive to contrast Weyl’s modifications of the Riemann theory with the theory that we have developed here:[21] Weyl: Distant comparison neither of vector magnitudes at a distance, nor of di- rections Riemann: Distant comparison of vector magnitudes, but not of directions Present theory: Distant comparison of vector magnitudes and directions.[22] 217. To Fritz Haber [Berlin,] 7 June 1928 Dear Haber, Dr. Robert Kauffmann, who sits on the board of the Berlin Electricity Works, A.G., an old acquaintance of mine and a zealous member of the Democratic Party,[1] has a plan for a propaganda effort in the service of peace that is to be con- cerned with the use of poison gas in war. I am asking you to promptly grant this man, who will telephone you, an interview. I don’t know your views on this subject,[2] but the man is smart, practical, and energetic. Health-wise, I am still no better. In about two weeks I am traveling to the Baltic in order to spend some time outdoors.[3] In the last few weeks I have found an en- tirely new way of realizing the general theory of relativity that may be ground- breaking.[4] Hoping that you are in good health, warm regards, your    g        g  g  g        hd a J 0 = h a