D O C . 9 1 R E L AT I V I T Y & E Q U AT I O N S O F M O T I O N 1 0 1 § 3. Second Approximation. Introducing the abbreviations , (16) then (8) takes on the form , (17) where means the operator . Furthermore, we note that: When the equa- tions (17) are fulfilled, then they remain fulfilled if one adds the expressions to the , where the four functions are chosen arbitrarily. From this, one can readily see that the can be normalized in such a way that the relations (18) are satisfied. Instead of (17), one then obtains . (19) This “separability” of (17) into (18) and (19), which is closely related to the diver- gence properties of the tensor ,[28] is of decisive importance to our in- vestigation. Namely, (18) and (19) contain a sort of “overdetermination” of the [29] in such a way that fulfilling (19) does not necessarily lead to fulfilling (18), so that equations (18) still contain an additional condition, whose possibility to be fulfilled is not self-evident.[30] We introduce the abbreviation . (20) Then, (19) takes on the form: . (21) [p. 241] g i 1 2 -- i g   i  – g i  x   x   i    i x   x       + + – –2Q i =  2 x2   x  i i x   + g i  i g i  x   i 0 = g i 2Q i = R i 1 2 g i R – g i S i Q i 1 2  i Q  – =  i 2S i =