Yashar Hezaveh had heard the rumours. Now, taking a seat in the second row, he was about to learn whether they were true.
Astronomers like him filled the room. It was a cold and gloomy January day in 2015 in Seattle, and they were attending one of the world's largest astronomy conferences. They wanted a peek at the first images taken with the new ALMA telescope, an array of 66 radio dishes arranged 5,000m above sea level in Chile's Atacama Desert.
These first snapshots would show off the telescope's capabilities. One of those high-resolution pictures, Hezaveh had heard, would reveal SDP81, a galaxy 12 billion light years away, first seen as a smudge by the Herschel Space Observatory in 2010.
We don't see these dwarfs, but we know they're there because of gravity
The pressing question for Hezaveh, a researcher at Stanford University in the US, was the quality of the new images. For years, he had been using computers to simulate how galaxies like SDP81 would appear to ALMA, the Atacama Large Millimetre/submillimetre Array. And he needed to know whether his simulations were accurate.
If they were, then he and his team would be in business. The astronomers were hunting tiny dwarf galaxies that could be hidden in images like this one. Dwarfs are virtually invisible, but their gravity could leave an extremely subtle imprint. "We don't see these dwarfs, but we know they're there because of gravity – which is pretty cool if you ask me," says Tommaso Treu, an astronomer at the University of California, Los Angeles, in the US.
Knowing whether these dwarf galaxies exist – and if so, how many there are – is crucial for solving one of the biggest cosmic conundrums around: what is the stuff that makes up the Universe? Also at stake is perhaps the best theory yet of how the Universe came to be. Despite decades of success, the theory still has question marks. And dwarf galaxies may hold the answers.
When you think of galaxies, you probably envision something like our own Milky Way: a spiralling disk of several hundred billion stars: a grand, celestial pinwheel.
Dwarf galaxies, on the other hand, are different beasts. They are diminutive and dim, sometimes only a diffuse gaggle of stars.
The stars are enveloped in a big blob of dark matter
While dwarf galaxies can be isolated, floating in space alone, they are also thought to hide around their bigger, more majestic counterparts, orbiting them like moons around planets.
The Milky Way, for instance, has nearly 50 such companions. Also called satellite galaxies, they are relatively tiny, weighing roughly 10,000 times less than the Milky Way, with some containing only a few thousand stars.
Like in all galaxies, the stars are enveloped in a big blob of dark matter, the unknown stuff that comprises around a quarter of the Universe. The dark matter accounts for most of the dwarf galaxy's mass, forming the gravitational glue that binds the galaxy together.
The biggest of the Milky Way's satellites are the Large and Small Magellanic Clouds, which appear as two fuzzy patches to the naked eye. The other satellite galaxies, however, are small and faint, detectable only with telescopes.
Astronomers found a handful by the mid-20th Century, but the majority remained unseen until the last decade, when modern astronomical surveys discovered dozens.
Cold dark matter fundamentally is a theory that says little things formed first
This was a remarkable revelation. In the late 1990s, before so many were found, astronomers were in a crisis, Treu says. Computer simulations were revealing a glaring discrepancy between how many satellite galaxies astronomers saw and what was predicted by their prized theory of cosmology: cold dark matter (CDM).
Dark matter, the theory posits, is cold: its particles move around space slowly (in this case, much slower than light). Because they were not zipping around too fast in the early Universe, their mutual gravity could corral them into small, dense clumps. These clumps eventually attracted the gas needed to form stars and become dwarf galaxies. Over time, these dwarfs merged together and became big galaxies like the Milky Way.
"Cold dark matter fundamentally is a theory that says little things formed first," says James Bullock, an astrophysicist at the University of California, Irvine. "They merged together to form bigger things over time."
If that is the case, you would have lots of tiny clumps and a fewer number of large clumps. In fact, the smaller the clumps, the more abundant they should be.
So far, the theory has been very successful at predicting the correct number and masses of the big clumps: big galaxies and clusters of galaxies. But astronomers have yet to corroborate the theory's predictions about the small clumps and dwarf galaxies.
We haven't found the thousands or hundreds of thousands that we need to find
In particular, thousands or even hundreds of thousands of dwarf galaxies could be around bigger galaxies like the Milky Way, circling them before merging. Around other distant galaxies, these dwarfs would be too small and faint for any telescope. But even around the Milky Way, astronomers had only found a handful, a discrepancy that became known as the missing satellite problem.
Over the last two decades, the gap between theory and observations narrowed when surveys started locating more satellites. But while the crisis abated, it did not go away.
"Some people were saying: problem solved. We don't need to worry about this; [the satellites] exist, we just need to know how to find them," Treu says. The problem was, "we haven't found the thousands or hundreds of thousands that we need to find."
If the satellites do not exist, then CDM could be wrong, or at least incomplete. "That's why the prediction is so powerful, why finding them will be so important," Bullock says. "You're really testing what cold dark matter means."
Another possibility, one that does not require upending or changing CDM, is that the satellites are indeed out there, but they do not have enough stars to be seen. After all, the theory only predicts the existence of small, dark and starless matter clumps called subhalos.
The galaxy's light was being distorted by the gravity of a closer, foreground galaxy
Dwarf galaxies are simply subhalos with stars, but how many stars – if any – formed in those clumps is another complicated problem altogether.
The most direct way to test this prediction of CDM is to go after the clumps themselves, regardless of whether they have stars. The only way to find these dark blobs is by measuring the influence their gravity has on their surroundings.
Around the Milky Way, for example, a thin stream of stars being yanked from another satellite galaxy might indicate the presence of a nearby subhalo or satellite galaxy. But such streams are so wispy and faint, they are difficult to observe.
To better gauge the existence of subhalos and dwarfs, astronomers must look for them around other galaxies. In which case, Hezaveh says, the only possible method will be to tease apart how their gravity bends light.
After waiting with anticipation, Hezaveh finally saw what he came to see. Just as the rumours said, ALMA had taken an image of SDP81, revealing two bright arcs that looked almost like flames.
The image was only visible for a few seconds
The galaxy's light was being distorted by the gravity of a closer, foreground galaxy. Acting like a giant lens, the gravity brightened SDP81, warped its appearance into an arc, and, in this case, split the image of the galaxy into two.
The resolution of the new image was as good as in Hezaveh's simulations. "This was incredible," he recalls. "That was a really powerful moment for me."
The image was only visible for a few seconds, just long enough for him to snap a picture. Right away, he emailed the photo to his colleagues.
Hezaveh knew that with data this good, he would be able to find an elusive subhalo. If one existed around the foreground galaxy, then its gravity would add another layer of distortion, like a water droplet on a lens. But these distortions are so subtle they require sophisticated calculations to pick out.
The foreground galaxy has no dust, which makes it invisible to ALMA
Over the past year, Hezaveh and his colleagues used these techniques to analyse SDP81. By comparing the observations with their simulations, the researchers could deduce the existence of a subhalo and its mass. Lo and behold, they discovered a subhalo weighing about a billion suns.
"It's a huge technical, computational, and scientific achievement," says Treu, who was not part of the team.
It is not the first time astronomers have used gravitational lensing to find subhalos. But what is different about this latest discovery is that it used the new and powerful ALMA telescope, which detects light with wavelengths longer than infrared. This radiation emanates from the warm dust that fills the background galaxy of SDP81.
The foreground galaxy, however, has no dust, which makes it invisible to ALMA. Without foreground glare, astronomers can analyse the background galaxy in more detail, making measurements with unprecedented sensitivity.
There are competing theories of dark matter
Crucially, the researchers could use their method to detect other, smaller subhalos; the yet-to-be-discovered lightweights that CDM predicts.
In particular, the researchers showed they could, in principle, find subhalos with a mass of only 10 million times that of the Sun, 100 times lighter than the one they found. "This is the regime where things start to get interesting," Treu says.
That's because there are competing theories of dark matter that make very different predictions about how many of these lower-mass subhalos exist, Hezaveh says.
For example, say dark matter is warm instead of cold. By definition, that means the dark matter particles would have been flying around faster in the early Universe, too restless to collapse into small clumps. By measuring the abundance of small clumps today, astronomers can deduce whether dark matter is warm or cold.
"If you find a lot of these dark clumps, you would put the nail in the coffin of warm dark matter, that's for sure," Bullock says. "It would really suggest something like cold dark matter is the truth."
It's an incredible piece of work, and it's finally starting to pay off. This is only the beginning.
But to find more subhalos, astronomers need more gravitationally-lensed galaxies. So far, though, the search using gravitational lensing had stagnated, turning up only a couple of subhalos. Now things may change.
"What we're excited about here is that this is the first lens that we studied and it happens to be exactly as sensitive as we thought," Hezaveh says. The galaxy itself also revealed a clumpy structure, providing the textured image that Hezaveh needed to discern tiny distortions from a low-mass subhalo.
"The moment I saw those amazingly thin and extended arcs, and tiny clumps in the source, I knew something very exciting was going to come out of it," he recalls. And if this first one was that good, then others might be just as sensitive, or maybe more.
"It's an incredible piece of work, and it's finally starting to pay off," Treu says. "This is only the beginning."
Just a few years ago, astronomers only knew of about 150 gravitational lens systems, but the number is rapidly rising. In the last couple of years, Hezaveh and his colleagues have used the South Pole Telescope to discover about 150 more, with hundreds more expected from other telescopes.
ALMA, meanwhile, is best suited for detailed, follow-up observations. So Hezaveh's team wants to point the telescope at these newfound systems and search for more subhalos. By analysing more galaxies over the next few years, he hopes they will have enough data to resolve the missing satellite problem and better understand the mystery of dark matter.
"Now is going to be the golden age of this approach," Bullock says. "The next few years should be pretty exciting."
Join over five million BBC Earth fans by liking us on Facebook, or follow us on Twitter and Instagram.
If you liked this story, sign up for the weekly bbc.com features newsletter called "If You Only Read 6 Things This Week". A handpicked selection of stories from BBC Future, Earth, Culture, Capital, Travel and Autos, delivered to your inbox every Friday.