A map of the most distant parts of the Universe – those objects that are at least 4 billion light years away – has been constructed using data from the Sloan Digital Sky Survey. Researchers have calculated how smoothly or clumpy the distribution of galaxies is at this scale, and have come to a surprising conclusion: the Universe is much more clumpy than models predict.
Matter in the present-day Universe is obviously arranged into clumps - stars, galaxies and galaxy clusters. But one would expect that the Universe originally started off in a state of relative smoothness. Therefore, one would expect to observe smoothness in those parts of the Universe which are many billions of light years away, since we observe those distant parts in the state that they were in billions of years ago due to the time it has taken for the light from them to reach us.
The very early Universe is expected to have been smooth because the Big Bang theory says that the Universe at that time was very small, so that light signals and matter could be exchanged between regions without much time delay. Therefore variations in temperature and density would have been smoothed out by exchange of heat energy and matter.
During inflation, the Universe expanded rapidly, with distances between some points increasing faster than light could cross them. These regions therefore became isolated from each other, no longer able to exchange matter and energy. Any fluctuations in the density of the Universe at this time became locked in, fixed for all eternity.
Over time, small peaks in the matter density became more pronounced, as the denser regions exerted a gravitational pull to draw further matter towards them. This is how objects such as stars, galaxies, and huge galaxy clusters formed, leaving huge voids of empty space between them.
The problem is that standard models of cosmology predict that the distant regions that have recently been mapped would have a clumpiness that varies only by about 1% over a length scale of 2 billion light scales. The observed variation is almost double that.
This means either that there is something wrong with the cosmological models used to make the predictions - possibly the assumptions about the distribution of dark energy, or the existence of dark energy at all - or that Einstein's theory of general relativity doesn't work on such large scales. Despite the success of general relativity at predicting the movements of planets, it could be that it is only an approximation to the true theory of gravity, in the same way that Newton's theory of gravity is only an approximation to general relativity.
There is of course a third possibility, which is that the unexpected result is actually due to systematic errors in the data, such as dust in our own galaxy blocking the view of more distant objects, or nearby stars being misidentified as distant galaxies. A further study, using data that the Dark Energy Survey will begin collecting in October 2011, will hopefully confirm or rule out this possibility.