If light from billions of years ago were trapped within a perfectly reflecting mirrored box, will it be redshifted in a way comparable to light from galaxies billions of light-years away? I had made up this thought experiment as a supposed illustration of how the redshift in distant galaxies is not due to them moving away but rather because of the expansion of the universe itself, but I worry that it doesn't work out the way I think it would. (Although I should have probably mentioned that the box should be very far away from any significant concentrations of mass.) I can only think of two real reasons why light trapped in such a box won't be redshifted: Even in principle, the gravity from the box itself or even the photons themselves would be sufficient to locally cancel out any expansion. Even with a sufficiently large box, say a hundred kilometers on each side, quantum mechanics will prevent any significant shifting in the frequency of the trapped light. I don't think that the couple of possible objections above would really apply, but I'm not sure (I'd probably have to spend a year absorbing the background and working it out). How does everyone else think the thought experiment would work out?