One of the things we don’t emphasise enough when writing about science is that most theoretical proposals about nature, made in advance of experiment, are wrong. Nature has a way of surprising us, so it’s vitally important to realise that science cannot proceed by pure thought and logic alone. Without the guidance of experiment, scientists are like those who choose to immerse themselves in sensory deprivation tanks. They inevitably tend to veer off into hallucination. So it is all the more telling when new unexpected clues about nature come from observations, or when some of our creative imaginings turn out to reflect the way nature really works.

ONE OF THE EXOTIC surprises physics has thrown at us in the past 30 years is the concept of dark matter. Dark matter was first inferred by Swiss astronomer Fritz Zwicky in 1933 when trying to figure out what stopped galaxies in massive clusters from flying apart. When I was a student, in the 1970s, the case for dark matter was suggestive, but not compelling. As a particle physicist, I became fascinated by the idea that particles could naturally be produced in the early universe with the necessary abundance and characteristics of dark matter today. But the game was wide open.

Neutrinos, which after all are known to exist, still seemed like a good bet to be dark matter particles, but it was still possible that our estimates about the amount of dark matter were wrong, and maybe it could be made up of boring things that simply don’t shine as brightly as stars or hot gas, such as planets.

Now, there is compelling evidence that cold dark matter is real; from direct measures of mass in galaxies and clusters, calculations of the abundance of light elements produced in the Big Bang, measurements of large-scale structure coming from millions of galaxies, and remarkable observations of the primordial seeds of structure observed in the cosmic microwave background (CMB) radiation from the Big Bang.

Observations of colliding galaxies suggest very convincingly not only that dark matter exists in profusion; but also that it cannot be made of normal matter. Moreover, candidates for dark matter may be on the verge of detection by direct detectors of the sort that I and others proposed more than 25 years ago, or perhaps may be produced in the coming years at the Large Hadron Collider.