The long list of unanswered questions about black holes contains one particularly surprising item: How do they eat? Unlike many of the riddles that black holes pose, this one seems so simple: What do you mean we don’t know how things fall into a black hole?

The question makes sense when you stop thinking about black holes as cosmic vacuum cleaners—which they are not—and start seeing them as astrophysical objects that (in many cases) play by the same orbital rules as stars and planets. The earth doesn’t fall into the sun. Likewise, all things being equal, an object orbiting a black hole at a safe distance should keep sailing peacefully and indefinitely along its path. And yet we know stuff falls into black holes all the time because as it does, its gravitational energy is converted to electromagnetic radiation—light—and the black hole shines.

Here’s another way to think of the problem. “Black holes have this intense gravity, but they’re trying to squeeze everything into this small volume,” says Shep Doeleman, director of the Event Horizon Telescope (EHT). “There are not very many ways to do that. You compress gas and it heats up. Take all this gas and it’s streaming toward the event horizon [the boundary of a black hole] and getting closer and closer together, it gets hotter and hotter and wants to fly away. Convincing all that gas to go through the event horizon is not a straightforward thing.”

Astronomers with the EHT published results today in Science showing that magnetic fields are essential to this process. In observations conducted in 2013, the astronomers found ordered magnetic fields near the event horizon of a black hole called Sagittarius A*, the 4-million-solar-mass giant at the center of the Milky Way.

Here’s how the process seems to work. As matter spins around a black hole it forms a flattened pancake, called an accretion disk. This gas, like pretty much all gas in the universe, is magnetized, and as the accretion disk forms, magnetic field lines thread through it. But an accretion disk is an intense place, where gas reaches billions of degrees and orbits at close to the speed of light. “The magnetic fields are dancing around as the accretion disk is torn apart by differential rotations,” says Michael Johnson, an EHT postdoc at the Harvard-Smithsonian Center for Astrophysics and lead author on the new paper. This stretching and dancing creates turbulence in the accretion disc, which generates friction—and friction drags down matter that otherwise would continue orbiting uninterrupted. Through this process, the accretion disk becomes a vortex of doomed matter bound to drain from our universe.

This basic picture has been around in various forms since the mid-1970s, but until the Event Horizon Telescope trained three radio telescopes on the galactic center, these magnetic fields had never been directly observed.

The EHT uses a technique called very long baseline interferometry, in which scientists collect data at geographically distant telescopes and then combine it later, thereby mimicking the action of a much larger telescope. Each of the stations participating in this observation—the Submillimeter Array (SMA) and the James Clerk Maxwell Telescope (JCMT) on Mauna Kea*; the Combined Array for Millimeter-Wave Research (CARMA) in California; and the SMT telescope at the Arizona Radio Observatory—were equipped to observe two polarizations of light. Through a complicated process of cross-correlation, the astronomers isolated light with a telltale polarization signature—evidence that strong magnetic fields, some highly ordered and some chaotic, are present in the immediate environment of the black hole, where that light originated. And that is a big first—the direct detection of the long-postulated magnetic fields at the heart of some of the biggest mysteries about astrophysical black holes.

It’ll be fascinating to see how theorists interpret these observations. And it’ll be even more interesting to see what kinds of results—including detailed maps and images—the EHT reels in as it expands from the three-station array used in 2013 to an eight-station-plus Earth-spanning network of telescopes. More on those efforts in another post.

*In the original version of this post, I neglected to mention the participation of the James Clerk Maxwell Telescope (JCMT). My apologies. The omission is doubly bad because only by using both the Submillimeter Array and the JCMT—each telescope recording a single polarization—were the astronomers on Mauna Kea able to observe both polarizations. So: although it's not true that each telescope that participated in this experiment recorded two polarizations, it is true that each station recorded both.