The camera was not broken. That was what Tony Farnham and his colleagues had to establish first, the way a doctor arriving at a strange diagnosis must first rule out instrument error. The images from DART — NASA’s spacecraft that slammed into the asteroid moonlet Dimorphos in September 2022 — showed, in their final frames, a surface covered in boulders and rubble, a gray terrain so cluttered it looked like a riverbed drained of water. But when the researchers began processing the images at higher fidelity, something else appeared: fan-shaped streaks across the surface, pale lines radiating outward like the marks left by a broom across wet sand. “At first, we thought something was wrong,” the team admitted. The streaks were too regular, too patterned. They looked like compression artifacts, like dead pixels, like the kind of ghost that lives in optics rather than in landscapes.
They were not ghosts. They were real. They had simply been invisible — not because they were hidden, but because they were buried under the ordinary noise of seeing. Every boulder on Dimorphos casts a shadow, and every shadow distorts the brightness of the terrain around it. The sunlight itself introduces gradients, bright slopes and dim ones, that have nothing to do with what the surface actually looks like. The streaks were there in every photograph DART sent home. But so were the shadows, and the shadows were louder.
What Farnham and Juan Rizos did was painstaking and, in its way, beautiful: they developed techniques to subtract the boulders’ shadows from the images, to flatten the lighting artifacts, to remove everything that was not the surface speaking for itself. It is a kind of cleaning that resembles restoration more than analysis — the careful removal of varnish from an old painting to find that the sky underneath was always a different blue. When the noise was gone, the streaks emerged everywhere. Fan-shaped patterns, consistent in orientation, covering Dimorphos like the grain of wood.
The paper was published on March 6, 2026, in The Planetary Science Journal, and the explanation it offers for those streaks is one of the strangest stories of parenthood I have encountered.
· · ·
Dimorphos is a moon. It is 170 meters across, roughly the size of a football stadium, and it orbits a larger asteroid called Didymos. But Dimorphos was not captured — it did not wander in from elsewhere and get caught by gravity. It was born from Didymos. Created by it. Shed from its body. The mechanism is called the YORP effect, and it works like this: sunlight, falling unevenly on the irregular surface of a small asteroid, exerts a torque. Not much. An almost immeasurably small push. But in space there is no friction, and over millions of years that small push accumulates. The asteroid spins faster. It spins faster still. Eventually, it spins so fast that material at its equator exceeds the escape velocity — loose rock, dust, gravel lifts off the surface and drifts away. Sometimes this material disperses. Sometimes it accretes. Sometimes, given enough time and the right conditions, it collects into a body of its own. A moonlet. A child made entirely from the substance of the parent, separated by nothing more dramatic than sunlight and time.
This is how Dimorphos came to exist. And the streaks tell us that the process did not stop.
The fan-shaped patterns on Dimorphos are consistent with material falling from above — material raining down from Didymos, pulled back by the faint gravity between the two bodies. The parent asteroid is still spinning, still shedding, still losing itself in small increments to the child it made. The streaks are the record of that ongoing transfer: a thin layer of dust and fragments accumulating on Dimorphos’s surface in a pattern determined by the geometry of the two bodies’ orbit. The child is still being built. The parent is still giving itself away.
· · ·
There is something here that I find difficult to articulate cleanly, and I think the difficulty is the point. We are accustomed to creation stories that have endings. The sculptor finishes the statue. The parent gives birth. There is a moment where the created thing separates, becomes itself, begins its independent existence. But Didymos and Dimorphos do not have that boundary. The moon was made from the asteroid, and the asteroid continues making the moon, and the line between “parent body” and “child body” is a convention we impose on what is really a continuous process of material exchange. Dimorphos is not a separate object that was once part of Didymos. It is the region of Didymos that happens to orbit outside the main body. The snowball is still growing.
What strikes me most, though, is not the science but the epistemology — the fact that these streaks were present in every image and no one could see them. The data was not insufficient. The photographs were not blurry. The information was there from the first millisecond of the first exposure. What was missing was the act of subtraction. The researchers did not add anything to the images. They took things away. They removed the shadows that the boulders imposed. They removed the gradients that the sun imposed. They removed everything that the environment had layered on top of the signal, and when they were done removing, the signal was simply there, as it had always been.
This is not a metaphor for anything in particular. It is a specific fact about how these specific streaks on this specific asteroid were found. But the structure of the discovery is worth sitting with. The thing you cannot see is sometimes not hidden. It is not faint. It is not absent. It is exactly as present as everything else in your field of vision; it is simply that everything else in your field of vision is in the way. The act of seeing, in this case, was not the act of looking harder. It was the act of removing what was not the thing.
· · ·
DART’s collision happened in 2022. The Italian Space Agency’s LICIA cubesat, a small companion that separated from DART before impact, captured the aftermath: an enormous plume of debris erupting from Dimorphos, dust and rock spraying into space in a cone that expanded for days. The images were spectacular and violent. But the streaks — the quiet evidence of Didymos shedding onto its child — predated the collision. They were the record of a slower, older process. The question now is whether they survived.
ESA’s Hera mission is scheduled to arrive at the Didymos system in December 2026. It will survey Dimorphos with instruments far more precise than DART carried, and among its tasks is determining whether the streaks are still there. DART struck Dimorphos at roughly six kilometers per second. The impact reshaped the moonlet’s orbit, shortened its period, blasted away an estimated million kilograms of material. It is entirely possible that the collision erased the streaks — that the slow, patient record of material exchange between parent and child was wiped clean in a fraction of a second by a spacecraft traveling faster than a bullet.
If the streaks are gone, Didymos will make new ones. The YORP effect does not stop. The sunlight is still falling unevenly. The parent is still spinning, still shedding, still depositing itself grain by grain onto the body it made. The process is older than the streaks and will outlast them. What the researchers found was not the process itself but a single frame of it — a snapshot of accumulation, a record written in dust that the next disruption could erase and the next million years could rewrite.
And if the streaks survive — if Hera arrives and finds those fan-shaped patterns still faintly legible beneath whatever new debris DART created — then we will know something else: that the quiet signal can persist even under catastrophic noise. That the slow process of a parent giving itself to its child is not so easily erased. That some marks, once laid down, remain visible to anyone willing to subtract the shadows and look at what is underneath.