The Milky Way’s Star Birth Recipe Is Not Universal

Sort the stars of the open cluster NGC 6067 by mass and count them, and the tally does not fall away smoothly. It slopes down gently through the lighter stars, then bends and drops faster, and the bend sits at about 1.23 times the mass of the sun. The description of stellar birth that most astronomers work with puts that bend at half a solar mass. These are raw counts, before any correction has been applied to them.

The bend has a name and a job. Astronomers call the mass distribution of newly formed stars the initial mass function, and because the small, dim stars in a distant galaxy can’t be seen individually, that function is what tells them how many invisible faint stars to credit for every bright one they can actually count. Get it wrong and every galaxy mass in the catalog is wrong with it. “One of astronomy’s basic assumptions may be oversimplified,” said Charles Steinhardt, an astronomy professor at the University of Missouri and an author of the study.

A cluster does not keep the population it was born with. Its heaviest stars burn through their fuel first and disappear from the top of the count, which is how the cluster’s age gets measured in the first place. At the same time the Milky Way’s gravity works on the cluster’s outskirts for hundreds of millions of years, and the lightest members, which drift outward, get pulled loose first. Both processes tilt the count. Neither one relocates the bend.

The reason is arithmetic. The standard form describes the distribution as two straight lines on a logarithmic plot meeting at a break. Stripping stars away multiplies the whole distribution by a smooth curve, which changes the tilt of both lines and leaves the meeting point exactly where it was. That is a property of this particular shape, and it is why the researchers chose it.

So the break is a birthmark. If two clusters show breaks at different masses, they were assembled from different mixes of stars to begin with, and no amount of subsequent aging or unraveling accounts for it.

Older Star Clusters Break at Higher Masses

The work was done in Mizzou’s College of Arts and Science by Steinhardt and Carter Meyerhoff, an undergraduate researcher, with Alexander Luening of the University of Rochester. They started from a catalog of 7167 clusters pulled out of Gaia’s third data release by a separate group in Heidelberg, and narrowed it hard. A usable cluster needs enough faint members detected below the break to anchor one side of the bend, which only 417 of them had, and enough of the rare heavy stars above it to anchor the other, which cut the working sample to 110. Across those 110, break mass rose with age.

The trend came with a trap built into it. Distant clusters are only measurable at all if their break masses are high, since a low break would sit below what Gaia can see at that range, so the sample picks up a spurious link between break mass and distance. The researchers modeled the distance effect and subtracted it. The age relationship survived and got stronger, which is what a real signal does when a contaminating one is removed.

Steinhardt, the paper’s first author, framed the consequence for other galaxies bluntly. Other galaxies, he said, “weren’t breaking the laws of physics,” and “we were measuring them with the wrong yardstick.”

Hotter Gas Favors Heavier Newborn Stars

Stars condense out of cold molecular gas, and the size of the pieces that condense is decided by a contest. Gravity pulls a clump inward. The gas pressure inside pushes back, and the speed of sound in that gas is the rate at which the push can travel across the clump and resist the squeeze. Warm the gas and sound moves faster, so pressure reaches farther before gravity wins, and the smallest lump capable of collapsing gets heavier.

Fewer small stars come out the other side, and the bend in the count slides upward. Older clusters condensed when the galaxy’s gas ran hotter and carried fewer of the heavy elements that let a cloud shed warmth, which points the predicted break masses in exactly the direction the measurements go. A star count of roughly 93,000 small stars near the sun reported the same drift in 2023, with the oldest local populations holding proportionally fewer low-mass stars than the standard recipe expects.

Why Earlier Studies Blamed Gravity

Several previous studies had noticed that older clusters break at heavier masses and concluded the opposite thing: that stripping had done it. They were not being careless. Those studies fit a different, curved formula to the data, and for that shape depletion does drag the characteristic mass upward, so an old cluster’s shifted peak could be pinned on gravity with no birth difference required. Older clusters are also more depleted simply because they have had longer, which tangles the two explanations together. The broken form untangles them.

Some of the ground underneath the measurement is still soft. Star counts have to be corrected for pairs too close for Gaia to separate, which the instrument logs as a single brighter object, and how to make that correction remains an open argument among the people who study it. The stellar models used to translate brightness into mass may also run biased at the low end. Either would move where the break falls, though both would move it in every cluster at once rather than manufacturing a trend with age.

None of this argues for throwing the recipe out. What the researchers propose instead is a family of recipes indexed to the conditions a batch of stars formed in. “The pattern we found is surprisingly clean,” Meyerhoff said. “Instead of applying the same model to every galaxy, astronomers could account for the conditions under which stars formed and select the IMF that best matches that environment.”

A Changing Star-Birth Recipe Could Alter Early Galaxy Masses

The clusters here span roughly ten million to a billion years, a thin slice of galactic history read from one galaxy. But the assumption being tested is applied everywhere, and it fails hardest where conditions differ most from the modern Milky Way. That means the early universe. Among the specific cases the paper points to are the galaxies that looked improbably massive in JWST’s first deep images, six of them apparently already grown to more than ten billion suns within 700 million years of the Big Bang. If those galaxies made proportionally fewer faint stars than the standard recipe assumes, their light was being converted into too much mass, and the sizes come down. “We’ve found that the universe is more complicated than we assumed,” Steinhardt said. “But we’re also getting closer to measuring it correctly.”

Gaia will not be settling this itself. The spacecraft finished its observations and has been shut down and passivated, with its fourth data release scheduled for December 2026. Whether the tilt of the count varies too, and not just the bend, is the harder question, and open clusters may be the wrong place to ask it: gravity does move the tilt. Answering it would take clusters whose unraveling can be modeled so precisely that whatever difference in tilt is left over has to have been there at birth. Nobody has that set of clusters yet.

FAQ