The Object That Came Back In 2002 Was Painted White. Back-Integration Says Apollo 12.
A body captured into Earth orbit for eleven months returned a spectrum matching titanium-dioxide paint over aluminium. Two independent back-integrations put it beside Earth in November 1969. Neither team will call it certain: 89 per cent, and 76.
HILO, Hawaii — The identifying measurement is forty minutes of a 2.2-metre telescope, taken on 5 October 2002, on a point of light of the eighteenth magnitude that was moving fast enough to need guiding by hand.
It looks like a paint chip.
The reflectance is nearly flat above 5,500 ångström and collapses below about 4,300 — the signature of titanium dioxide, the white pigment in nearly every coating sprayed on nearly every launch vehicle of the 1960s. Under it, the continuum behaves like aluminium.
The object was captured into a temporary orbit of the Earth in the spring of 2002, found by an amateur on 3 September of that year, and released back into orbit around the Sun in June 2003. It is roughly the size of a school bus. Its area-to-mass ratio, derived from the way sunlight pushes it about, is what an empty tank would give and not what a rock would.
Two runs, one week in 1969
Two teams have integrated its orbit backwards, twenty-two years apart, with different force models and no contact between them.
Both close on the same week. Both put the object beside the Earth in the second half of November 1969, which is when Apollo 12 left, and both identify its third stage as the most likely body.
Neither will say it is.
The 2004 run, 10,000 clones, gives 89 per cent. The 2019 run, 40,000 clones and a heavier treatment of solar radiation pressure on an empty tank, gives 76. This newspaper asked both teams for the covariance matrices and the code rather than the conclusions, re-ran each, and reproduced both to within half a percentage point.
The tails are alive. On the 2019 run, one clone in seven closes on an earlier Apollo stage — mostly Apollo 8, December 1968 — and one in twelve on a Soviet upper stage, of which two remain unexcluded. Two clones in a hundred reach the window and arrive at nothing at all.
What 89 per cent is not
It is not a fact. It is a very strong probability with a live remainder, published with its uncertainty by people who did not have to publish it that way.
That standard is rarer than it should be and it is the standard this desk keeps citing: two amateur trackers did the same thing with a Russian upper stage in 2009, publishing every step of their working, and that file sits in our vault at PX-2191.
Dr. Fenella Oyarzún-Whitmarsh, who took the 2002 spectrum, has objected in writing for twenty-four years to being cited as having identified anything. Her account of how a likelihood became an announcement — a circular, a wire service, three days — runs unedited at the foot of this brief and is the most useful thing on this page.
Her colleague Dr. Bertrand Kagame-Ellis, who took the only other spectrum in June 2003, agrees about the pigment and is milder about what it proves. Titanium dioxide in a silicone binder was nobody's monopoly, and the published reference library for Soviet stage coatings of that era runs to four samples, none of them flown.
The grating
Both integrations return the object to the Earth's neighbourhood between March 2040 and November 2041. Beyond the next close approach the forward solution degrades quickly, because sunlight pushes an empty tank according to an orientation nobody can predict.
No telescope time is scheduled. The time-allocation committee, asked in writing, points out that proposals cannot be filed more than one semester ahead — a fair answer, and also the reason.
The instrument is the harder problem. The grating that made the 2002 measurement was decommissioned in 2016. Its replacement begins at 4,200 ångström, on the wrong side of the only feature that identifies the paint.
Nobody suppressed anything. Somebody made a defensible procurement decision for the science done on that telescope every night, and one consequence, unnoticed for a decade, is that the measurement which named this object cannot currently be repeated where it was made.
What is in solar orbit is a discarded piece of a launch vehicle, painted white, almost certainly Apollo 12's, and it will come back when the two people who have ever looked at it closely are 85 and 82.
Dr. Oyarzún-Whitmarsh asks us to record that she has twice offered to write the observing proposal for whoever holds the time in 2040, that the offer is open, and that it stands for as long as she can hold a pen.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
Sources & Method
We did not ask either team which stage it is. We asked for the covariance matrix, the force model and the code, re-ran both integrations ourselves on rented hardware, and then put the tail of each distribution — not the peak — to the two people who have ever taken the object's spectrum.
Who we spoke to
- Dr. Fenella Oyarzún-Whitmarsh, Astronomer, 71; took the reflectance spectrum in October 2002. Interviewed by video call twice and sent the full draft, the figure and both re-runs; replied in writing at 1,900 words June and August 2026 Objects to being cited as certain, and objects to something else in this brief that we had not expected. Printed entire below.
- Dr. Bertrand Kagame-Ellis, Astronomer, 68; took the only other spectrum of the object, June 2003. Interviewed by telephone twice; supplied his reduced spectrum and his original observing log July 2026 Agrees the coating is titanium-dioxide white. Disagrees with his colleague about how far that narrows the field, and says so mildly.
- Dr. Sunniva Okorafor-Lindqvist, Dynamicist; author of the 2004 back-integration. Supplied the covariance matrix, the force model and the code on request, unconditionally, within two days May 2026 Our re-run reproduces her 89 per cent to within half a point.
- Prof. Casimir Nwachukwu-Vance, Dynamicist; author of the 2019 back-integration. Supplied covariance, code and the full 40,000-clone output; answered fourteen written questions May and June 2026 Thinks the 2004 radiation-pressure treatment is too light for an empty tank and has said so in print since 2019.
- The observatory's instrument sheet and time-allocation committee, Published grating specifications, and the committee's written answer to us. Asked in writing whether any proposal exists for the 2040 window; replied in nine days July 2026 The answer is none, and the committee points out that proposals cannot be filed more than one semester ahead, which is a fair answer and also the reason.
Documents
- PX-2191 — Twenty-six hours of published record, 9–10 December 2009, with two amateur trackers' working accepted
What we could not confirm
- Whether the paint distinguishes American hardware from Soviet. Titanium dioxide in a silicone binder was nobody's monopoly. Our two spectroscopists disagree about how far the measurement narrows the field, and the published reference library for Soviet stage coatings from that period runs to four samples, none of them flown.
- The 2040 window. Both runs put the next capture between March 2040 and November 2041, and both say the forward integration degrades quickly after the next close approach, because sunlight pushes an empty tank differently depending on an orientation nobody can predict. A twenty-month window is what the physics currently supports; a diary date is not.
- Whether the object is an intact stage. The area-to-mass ratio derived from its non-gravitational motion, about 0.008 square metres per kilogram, is what an empty tank would give. A large panel, or a stage that has come apart and stayed together in a cloud, would give something similar, and no image ever taken of it resolves more than a point.
How Others Covered This
The same events, as reported elsewhere on the same day. We list what each outlet had that we did not, as well as what we had that they did not — including where we come off worse. Why we print this.
-
The Meridian TelegraphIT CAME BACK: THE APOLLO ROCKET THAT ORBITED EARTH AGAIN — AND RETURNS IN 2040
Printed the identification as settled and the 2040 date as a diary entry.
Had that we did not
The paint chemistry, correctly, and the capture and release dates, correctly. Their science desk read the papers.
Left out
The eleven per cent, the twenty-four per cent, and the two written requests from the astronomer whose spectrum they were citing, asking them to use the word 'probably'.
-
The Continental WireObject In Temporary Earth Orbit Identified As Rocket Stage, Astronomers Say
Wrote up both integrations at length and got every number right, including both probabilities.
Had that we did not
The covariance discussion, which no other outlet attempted, and a clear account of why the two runs differ.
Left out
That the spectrograph which produced the identifying measurement was decommissioned in 2016 and its replacement does not reach the wavelengths where the paint signature lives. One call to the observatory establishes it.
-
They Buriedthis newspaperThe Object That Came Back In 2002 Was Painted White. Back-Integration Says Apollo 12.
Asked both teams for the covariance and the code rather than the answer, re-ran each, and reported the tail.
Had that we did not
Both distributions in full, the instrument's blue cutoff, and Dr. Oyarzún-Whitmarsh's objection to this brief at her own length.
Left out
Our headline says 'Says Apollo 12'. Our fourth paragraph says 89 per cent. Dr. Oyarzún-Whitmarsh told us before publication that the headline travels and the paragraph does not, and she has been right about that twice before, in this newspaper. We ran the headline. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Dr. Fenella Oyarzún-Whitmarsh sent the full draft, the figure and both of our re-runs on 4 August 2026, with three weeks and no word limit. Replied 18 August 2026, asking that it run whole. Printed unedited, including the last part, which is about us.
I have now spent twenty-four years being quoted as having identified a rocket, and I would like to use your space to say precisely what I did, because it is smaller than the sentence that has attached itself to my name.
I had forty minutes on a 2.2-metre telescope on a moving target of the eighteenth magnitude. I obtained a reflectance spectrum. It is nearly flat above 5,500 ångström and it falls off a cliff below about 4,300, which is the signature of titanium dioxide, and beneath that the continuum behaves like a metal. I wrote, in the circular, that the spectrum was consistent with a titanium-dioxide-bearing white paint over an aluminium substrate. Every word in that sentence was chosen. 'Consistent with' is not 'is'. 'A white paint' is not 'the white paint used on the S-IVB'. I did not know then and do not know now what was on the outside of every upper stage launched in that decade by anybody.
Within three days a wire service had 'astronomer identifies Apollo rocket'. Within a week it had lost the astronomer and simply had the rocket. I wrote to two outlets, politely, and received one courteous reply promising a clarification that never appeared, and one no reply at all. I did not write to the third because by then I was tired, and I want that on the record too, because part of how a likelihood becomes an announcement is that the person who made the measurement gets tired.
So. On the dynamics I defer to the dynamicists, and I note that both of them, unlike everybody who has written about them, publish their uncertainties. Eighty-nine per cent is a strong result. It is also one chance in nine of being wrong, and one chance in nine is not a rounding error, it is the sort of number a person should be able to hold in their head while reading a headline. Professor Nwachukwu-Vance gets three chances in thirteen. You have drawn both, honestly, and I am grateful, and I do not expect it to make the slightest difference to what is repeated.
Now the part you will not like.
You end your brief by observing that no telescope time is scheduled, that the instrument is gone, and that the two people who have ever taken a spectrum of this object are 71 and 68. You have made our ages into a closing cadence. I know exactly what that paragraph is doing, because I have read your newspaper for years and it is the move you make when you want a dull story to ache a little at the end.
Here is what is wrong with it. It is not sad that I am 71. I have had a good career and I am busier than I would like. What is actually the case is that this object is worth about four hours of telescope time once every four decades, and no proposal system on Earth is designed to want something in 2040 — you cannot apply for time fourteen years out, and any committee that granted it would be doing something irregular. That is not neglect. It is an allocation system doing what it was built to do, on a timescale it was not built for.
And the instrument is the honest part of your brief and you have put it in the middle instead of the end. The grating that made this measurement possible was decommissioned in 2016 and its replacement starts at 4,200 ångström, which is on the wrong side of the only feature that matters. Nobody suppressed anything. Somebody made a sensible procurement decision for the science that gets done every night, and a consequence of it is that the specific measurement which identified this object cannot currently be repeated on that telescope. Write that. It is duller than our birthdays and it is the actual finding.
One request, which is not a correction. If you must print my age, print alongside it that I have twice offered to write the observing proposal for whoever holds the time in 2040, that the offer is open, and that it will remain open for as long as I am able to hold a pen.
Published unedited under our right-of-reply guarantee.
How was this story?
We publish the result, whatever it is. Reader verdicts appear on the front page and in our newsroom metrics.
Readers' Letters 0
Printed at once under the name you give and read by the desk afterwards; anything unfit is removed, with a note saying so, and nothing else is ever deleted — only corrected. Letters that changed something in the story carry a mark saying so, and there are 5 of those across the archive.