This is written 23 Feb. 2007 after the return from a major trip abroad.
    A little later than our "second thoughts" about seeking to find an
analogue to the "Einstein-Maxwell equations" where our vacuum equation
for space-time would be the presumed proper vacuum equation in the absence
of gravitating matter, a little later than these "second thoughts", 
the idea occurred that the basic theme of Einstein's attempt to fuse 
the descriptions of the electro-magnetic and the space-time-gravitational
fields into one equation (for 16 variables) might be VERY NATURAL and
indeed more natural than the originally considered idea.
    (That "originally considered idea" was to seek an appropriate
natural generalization of the Einstein-Maxwell equations.)
    One nicely relating fact is that in a context of presumed flat
space-time the tensor describing an electro-magnetic field (which
could be thought of as infinitesimal, so as not to gravitate) itself
naturally satisfies the wave operator (or covariant d'Alembertian).
    If we begin with an Einstein-like combination of the electro-
magnetic field and the gij field of the original sort then in terms
of the NEW gij tensor we will have new versions of the Rij tensor and
others (as Einstein did).
    Then MAYBE our Lagrangian, dependent simply on Rij and R, could
be rederived or re-interpreted in terms of the new Rij, etc. And then
also a system of equations COULD derive as Euler-Lagrange equations
derived from that. Or a favorable FORMAL generalization might exist.
    In any case, it seems interesting to consider what sort of equations
for the electromagnetic field would be derived AND whether or not a
natural sort of gravitating action of the electromagnetic fields would
appear.
   ////////////////////////////////////////////////////////////////
    And if this WERE successful it (perhaps) would suggest that the
inertial and gravitational mass of matter is, ultimately, derived just
from associated electromagnetic fields. It has long been known that
the electron and the proton can be modeled as particles where the
electro-magnetic field is more or less concentrated and the MORE  
CONCENTRATED field of the proton corresponds to the much greater mass 
of that particle, compared with the mass of an electron.