It can serve as an example to scientific communities that are clinging to ageing infrastructure at the expense of the next generation.

On 10 January, the US Department of Energy (DOE) announced that the United States will close its only particle collider, the Tevatron, located at Fermilab in Batavia, Illinois. The decision to shutter the decades-old collider this year (see Nature 469, 141; 2011) goes against the recommendations of the US particle-physics community's scientific advisory group, which believes that the machine still has work to do. But given the financial constraints facing Fermilab, and its ambitions for the future, the choice is at once courageous and correct. Moreover, it should serve as an example to other scientific communities that are clinging to ageing infrastructure at the expense of the next generation.

As is often the case in high-energy physics, the success of the Tevatron can be measured in orders of magnitude. Since the Tevatron's opening in 1985, its luminosity, a measure of the number of collisions produced, has increased by a factor of 20 million. Its detectors have gathered data at an exponential rate, along the way discovering the top quark in 1995 and the tau neutrino in 2000.

But few discoveries have emerged from the accelerator in the past decade. The search for heavier particles, such as the Higgs boson, a critical part of the mechanism for mass, uses higher luminosities and energies than the Tevatron could ever produce. To make up for its lack of power, the machine has been forced into a war of attrition: running for years beyond its expected lifetime in the hope of collecting enough data to spot something new. It is an inefficient way to do science.

Despite its mounting obsolescence, scientists have continued to push to extend the Tevatron's run. In the beginning, the case was made that it could yet beat its European rival, the Large Hadron Collider (LHC) at CERN, Europe's particle-physics lab near Geneva, Switzerland, to the discovery of the Higgs. That argument held some weight while the European collider was facing early teething pains, but last year the LHC roared to life and is now churning out data at a staggering rate (see page 282). The latest scientific case for keeping the Tevatron running until the end of 2014 revolves around its ability only to plug gaps in the LHC's work.

Although scientifically persuasive, this argument must be measured against the US$35 million needed each year to operate the machine. That sum might seem small in comparison to the US research budget as a whole, but it represents roughly 10% of Fermilab's annual operating funds. Lab leaders would like to put the money elsewhere: faced with the truth that the Tevatron cannot run forever, they are hoping to develop a strong programme in neutrino physics. The new direction will keep the scientists at Fermilab working for years to come and could even lead to future collider projects.

To administrators at Fermilab and the DOE, the decision was clear: the lab must shut down the Tevatron in order to move forward. In one sense it is a risky strategy — to trade a steady income for the promise of better to come. Funds for the new programme are far from guaranteed and could yet be a tall order in the current economic climate. But in making the sacrifice, the lab showed that it is serious about its new neutrino programme. It is also providing its political supporters with evidence that it is acting as a responsible research body, something that may help them win the additional money needed for the new programme.

The move stands in stark contrast to other fields that have chosen to protect the old over the new. Some in the astronomy community have lobbied hard to keep 'historic' facilities such as the Arecibo Observatory in the United States and Jodrell Bank in the United Kingdom. Elsewhere in Europe, geriatric research reactors and synchrotron light sources are kept running more out of national pride than scientific need.

The case for closing many of these facilities is less clear-cut than the Tevatron, however. The collider has obviously been superseded by its European rival, and high-energy physicists have little choice but to move on. By contrast, telescopes and synchrotrons can always gather more data on another star or a new material. Yet at some stage, the scientific value of these facilities becomes so diminished that it must be weighed against the new opportunities the funding could give to the community. Rather than fighting protracted political battles against their inevitable closure, researchers should provide new ideas and facilities that could be leveraged using funds from the closure of the last generation. Such a strategy is more risky, but it will ultimately prove much more rewarding.