# why does the CMB change so slowly

If anisotropies of the CMB originated from "quantum fluctuations of matter in a very tiny space that expanded to the size of the observable universe we see today" (quote from Wikipedia) would they represent a cross section through some feature that is roughly spherical and if such a feature could be considered to be one light year across ( I appreciate that is a tricky statement) it should change dramatically in appearance over a year.

-
– dmckee Feb 24 '14 at 22:58
Note that at $z$ of 1000, 1 light year is a few milliarcseconds across. – Chris White Feb 25 '14 at 12:59

As you point out, each year the last-scattering surface moves another light-year back, but we should be looking at the same structures. As dmckee suggests in his linked post, we might expect the physical size of structures in the CMB to be roughly $$(\rm{angular\ size}) \times (\rm{radius\ of\ observable\ universe}) \sim (1^\circ) \times (13\times10^9\ ly) \sim 2\times10^8\ ly$$ Where $1^\circ$ roughly corresponds to the first peak in the CMB angular spectrum. Thus, the size of these structures is much much greater than $1\ \rm{ly}$, and so we shouldn't expect their appearance to change much over a year timescale.