Cosmologists have long puzzled over the origin of structure in the universe. In whichever direction of the sky you look, there is a hierarchy of stars, galaxies and galaxy clusters. Yet intuitively, the Big Bang should have created a uniform spread of matter and left the universe a rather uninteresting place.

The breakthrough came with the discovery of temperature fluctuations in the cosmic microwave background radiation, formed some 380 000 years after the Big Bang, which supported the idea that the early universe was a plasma filled with tiny fluctuations in density. The denser areas caused the matter to "clump" together, providing the seeds for the cosmic structure we see today.

Unfortunately, our theory of gravitation – Einstein's theory of general relativity – cannot account for the extent of clumping without invoking the right amount of a mysterious substance called "dark matter". Originally introduced in the 1930s to explain anomalous galaxy dynamics, dark matter (which cosmologists think could make up to 95% of matter in the universe) is gravitationally attractive yet does not couple to light. But crucially it can describe how the initial plasma fluctuations were sustained for long enough to allow large structures to form – on its own, general relativity attests that they simply petered out. Even though dark matter has never been observed, the majority of physicists now believe that general relativity combined with dark matter is the only satisfactory explanation for the universe's large-scale structure.

However, in recent years there has been growing support for alternative theories of gravitation to general relativity that do away with the need for dark matter altogether. One of these, devised by Jacob Bekenstein in 2004 at the Hebrew University of Jerusalem, uses vector and scalar fields in addition to the tensor used in relativity, hence the name "TeVeS" (Tensor Vector Scalar). TeVeS has already been shown to explain galaxy dynamics without the need for dark matter. But now, building upon the numerical studies by Pedro Ferreira and colleagues performed earlier this year, Scott Dodelson and Michele Liguori from Fermilab in the US have confirmed that TeVeS can also provide such sustained plasma fluctuations. (See figure: "Powering galaxy formation".)

In particular, the pair found that it is TeVeS's additional vector field that provides the "extra" gravitation on large mass scales, meaning that denser regions of the primordial plasma could have accreted matter much faster than general relativity alone (without dark matter) predicts.

"Their elegant calculation brings us closer to understanding why a modified theory of gravity may be a viable alternative to dark matter," Ferreira told Physics Web. "The presence of the cosmic [vector] field – akin to an aether – drives the growth of the structure, making this theory compatible with observations of the cosmic microwave background and the distribution of galaxies."