🚀 go-pugleaf

In article <1733@cvl.UUCP>, nelson@cvl.UUCP (Randal Nelson) writes:
> 
> The mention of hyperspherical models made me wonder how we know that
> the universe really is as extensive as the distance computed from
> highly red-shifted objects suggests.
> In particular, how do we know that the univers is not (relatively)
> small with a topology that permits closed geodesics.
> In this case, highly red-shifted light has just looped the loop
> a couple of times, which would allow the entire universe to be in
> communication.
> This would have interesting consequences. For instance, certain
> object would represent earlier stages of our own galaxy if we knew
> where to look.
> I have never heard this notion discussed.
> Does anyone know what (if anything) eliminates such a model??

One of the simple tricks that could allow this is just to divide
the usual model up into cubes and then require that all the
cubes be identified with each other.  This makes me slightly
uncomfortable, but I can`t think of any reason why it can
be ruled out.  Then the only limitation on the smallest size
that could be tolerated comes from local observations of
the distribution of clusters and galaxies.  I *think* these
require a "cell" size of at least a few hundred megaparsecs.
I remember John Ellis talked about this a few years ago at a
seminar here.  He may have published the idea in Nature, but
I really can't remember the reference.
-- 
"More Astronomy                Ethan Vishniac
 Less Sodomy"                  {charm,ut-sally,ut-ngp,noao}!utastro!ethan
  - from a poster seen         ethan@astro.AS.UTEXAS.EDU
    at an airport              Department of Astronomy
                               University of Texas