Le 03/11/2017 =C3=A0 04:03, Steve Willner a =C3=A9crit=C2=A0:
Good grief! Galaxies are nowhere near thermal equilibrium, and their
spectral energy distributions (SEDs) look nothing like blackbodies.
Most galaxy SEDs are double-peaked, and neither peak has a blackbody
SED. And even if galaxies had blackbody SEDs, a sum of blackbodies
at different temperatures (or different redshifts if all galaxies had
the same temperature, which they don't) does not give a blackbody
SED.

I am not speaking about any foreground galaxy of course.

If the sea of galaxies extends to infinity (or to huge distances) the
farther you look, the more galaxies you will observe for a given solid
angle. At great distances you will see a wall of galaxies that fills
completely the view. The (very red-shifted) light from those galaxies is
the CMB.

[[Mod. note --
As Steve Willner noted in this newsgroup a few days ago
(article ), adding up the light from
a bunch of galaxies (some redshifted) wouldn't produce the
very-close-to-black-body spectrum that the CMB has.

Moreover, you're basically invoking Olber's paradox here... but you
only get the every-line-of-sight-intersects-a-galaxy result *if*
(a) the universe has a flat topology (ok, we're probably close to
that), and
(b) there's no redshift to reduce the energy we receive from the
more distant galaxies, and
(c) the galaxies are infinitely old (so their light has had time to
get to us) *and* have been producing lots of light for that
infinite time).
The problem is that (c) completely contradicts everything we know
about the luminosity evolution of galaxies and their component stars.
-- jt]]