Some of the oldest galaxies in the Universe have three times more stellar mass, and so many more stars, than all current models of galaxy evolution predict.
The finding comes from the Atlas3D international team, led by an Oxford University scientist, who found a way to remove the 'halo' of dark matter that has clouded previous calculations.The team's analysis means that all current models, which assumed for decades that the light we observe from a galaxy can be used to infer its stellar mass, will have to be revised.
It also suggests that researchers have a new riddle to ponder: exactly how galaxies forming so early in the life of the Universe got to be massive so fast.
A report of the research is published in this week's Nature.
'The light we see from galaxies is just the tip of the iceberg, but what we really need to measure are galaxy masses that all models directly predict,' said Dr Michele Cappellari of Oxford University's Department of Physics, who led the work. 'Galaxies can contain huge numbers of small stars, planets or black holes that have lots of mass but give out very little or no light at all.
'Up until now models assumed that stellar light could be used to infer the stellar masses and any remaining discrepancy with the observed total mass could be hidden behind a "halo" of dark matter. Our analysis shows that they can't hide any longer: galaxies are diverse and some have many more stars and are even stranger than we'd assumed.'
Up to now the key limitation on what it was possible to say about the stellar mass of galaxies was the difficulty in separating this out from the mass contributed by dark matter. Various attempts from independent groups failed to provide a conclusive answer.
The new analysis succeeded thanks to the availability of two-dimensional maps of stellar motions for a large sample of galaxies, combined with sophisticated models.
By disentangling stellar mass from dark matter the team was able to show that instead of the relationship between observable light and stellar mass being universal, it varies between different types of galaxies - with some older galaxies having three times the mass suggested by the light they give off.
Dr Cappellari said: 'The question of how you should turn light from a galaxy into a prediction of its mass has been hotly debated but up until now nobody has been able to kill off the idea that there's a simple and universal way to convert observed light into mass'. We now think we've done that by eliminating the "fuzziness" in models caused by dark matter.
'It's exciting because it reveals how much more there is to discover about how galaxies, and the early Universe itself, evolved.'
This research is part of the Atlas3D project and is part-funded by the Science and Technology Facilities Council, the UK sponsors of astronomy and of the William Herschel Telescope (WHT) that was used by the team. Dr Michele Cappellari is supported by a Research Fellowship of the Royal Society.
A report of the research, 'A systematic variation of the stellar initial mass function in early-type galaxies', is published in Nature.