Piri Reis Drew A Southern Coast. It Ends Where The Hide Ends.
One sheet of gazelle hide, 90 by 63 centimetres. Fitted to Antarctic bedrock the famous coast misses by 340 kilometres; fitted to Brazil, rotated and squeezed, by 61. The squeeze begins 2.1 centimetres from the trim.
A months-long reporting project. Documents cited below are held in The Vault and available to readers.
ISTANBUL, Türkiye — The surviving fragment of the map Piri Reis signed in 1513 is one sheet of gazelle skin, 90 centimetres by 63, cut square down its right-hand edge with a knife.
This newspaper paid to have it examined as an object rather than as an argument.
A twelve-band multispectral mosaic of the whole sheet. Two historians of cartography commissioned to fit the same control points, separately, without contact. A GIS analyst sent nothing but coordinates and not told what they were.
The southern landmass everybody argues about carries 43 place names. All 43 appear on other charts of the Brazilian and Patagonian coast. Not one of them is a place anybody has proposed for Queen Maud Land.
Two fits, and 279 kilometres between them
The claim is that the coast along the bottom of the sheet is Antarctica drawn before the ice.
Dr. Aygün Merzifonlu-Rasch in Vienna and Professor Isolde Vanbrugh-Achterberg in Coimbra each took 61 identifiable points off the mosaic and fitted them against two targets: the subglacial bedrock surface of Queen Maud Land, and the coast of South America.
Against the bedrock, the best fit either of them could obtain gives a root-mean-square error of 340 kilometres.
Against South America — rotated 41 degrees and compressed along the coast run by a factor of 1.7 — the error is 61 kilometres, which is unremarkable for a chart of that century.
Their two answers differ by four kilometres. Neither knew the other was working.
Coen Okuda-Aldersey, who was told only that he was fitting one coastline to two candidates, returned the same pair of numbers to within a kilometre and asked in his covering note why the job had been described as hard.
The bend starts 2.1 centimetres from the edge
The compression is the interesting number, because it is not constant.
For the first 31 toponyms the coast runs at the same scale as the sheets it borrows names from, within four per cent. Then, across about three millimetres of parchment, the coastline turns 41 degrees east, and everything after the turn is squeezed to 59 per cent.
The turn begins 2.1 centimetres from the trimmed edge of the skin. The last twelve names are Patagonian names on every other chart that carries them.
A draughtsman working from portolan sheets onto a hide bends the coast to keep it on the hide. It is among the most ordinary things a sixteenth-century compiler does, and it is visible here because the trim survived and the rest of the map did not.
The nearest thing in our own vault, PX-2125, settled a fifty-year argument about the Bimini stone by looking at seventeen cores instead of the argument. This is the same move on a smaller object.
Four letters, not one
Every retelling of the Antarctic identification quotes one sentence from a US Air Force reply of 1965.
We asked for the file. There are four letters in it.
An enquiry of 6 July. The reply of 6 August from Lieutenant Colonel Harlan D. Ostrowski, then commanding the 12th Reconnaissance Technical Squadron, which contains the famous sentence and three hedges in the same paragraph. A second enquiry of 19 August, pressing him. And his answer of 2 September, which is the one nobody prints.
In it he says he has never seen the map, only a photograph of it; that he cannot speak to the projection and will not be drawn on it; and that he would prefer his August letter not to be quoted without the qualification attached.
Twenty-one of the twenty-two reprints we traced stop the quotation one sentence early.
The twenty-second is Alun Ferreira-Stavely's, and he printed the rest of the paragraph in an endnote in 1998, and told us so himself, and his reply is at the foot of this page. It is the most serious thing in this story. In it he says that in 1991 he cut the word apparently out of a serving officer's sentence for reasons of rhythm, that eleven other writers copied his version, and that Ostrowski wrote to him in 1979 and he did not reply.
What this desk takes from it
The measurement is dull and it holds. A coast fitted to bedrock misses by 340 kilometres and fitted to Brazil misses by 61, and the difference between those numbers is not a matter of interpretation.
What we cannot measure is the part Ferreira-Stavely puts to us, and he is right that we ducked it. The skin explains the geometry. It does not explain why a compiler with eight named source charts in front of him kept going onto ground he could not fit, when the careful thing was to stop at the last name that fitted.
Something made him want the extra twelve names on the sheet. Imaging cannot see wanting.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
Sources & Method
We did not argue with the identification. We paid to have the sheet measured: a twelve-band mosaic made under museum supervision, then two historians of cartography commissioned separately to fit the same 61 control points against the same two candidate coastlines, neither knowing the other existed, and a GIS analyst given the coordinates with the subject withheld. Then we went and found the 1965 letters instead of quoting the sentence everyone quotes.
Who we spoke to
- Dr. Aygün Merzifonlu-Rasch, Historian of cartography, Vienna; first cartometric fit. Commissioned to fit 61 control points against two candidate coastlines, with no sight of the second analyst's work and no contact with her February–May 2026 340 km against bedrock, 61 km against South America. Her working file is published entire.
- Prof. Isolde Vanbrugh-Achterberg, Historian of cartography, Coimbra; second fit, blind. Given the identical mosaic and the identical brief, and told before starting that both fits would be published whether or not they agreed March–May 2026 Differs from the first fit by 4 km. Disagrees about six of the toponym readings and says so at length.
- Coen Okuda-Aldersey, GIS analyst. Sent a list of 61 coordinate pairs and two candidate coastlines, with no map, no names and no indication of the subject May 2026 Returned both errors to within a kilometre of the historians' figures and asked in his covering note why the job had been described as difficult.
- Bahar Tunç-Levering, Imaging scientist. Made a twelve-band multispectral mosaic of the whole sheet under museum supervision; the raw bands are published with this story January 2026 Also measured the trimmed edge at 40× and reports a knife cut, not a tear, along its whole length.
- The 1965 correspondence file, Four letters, July–September 1965, with the squadron's outgoing log for the quarter. Obtained from the enquirer's papers in a university manuscript collection; copying fee paid, all four transcribed and published April 2026 The famous sentence is in the second letter. The fourth is the officer asking that it not be quoted without the qualification.
Documents
- PX-2125 — X-radiographs of seventeen oriented cores taken from a submerged carbonate formation, Bahamas, 1977, with the 1979 radiocarbon report accepted
What we could not confirm
- Whether the compression begins at the trim, or the trim was made where the compression already was. We can measure both to a millimetre and we cannot put them in order. If the sheet was cut after the drawing was finished, the bend needs a reason we have not recovered.
- Six of the forty-three toponym readings. Our two historians agree on thirty-seven and disagree on six, three of which are abraded past argument. All six sit in the compressed run, which is where the reading matters most, and we have printed both readings side by side rather than choosing.
- The fifth letter. The squadron's outgoing log for that quarter lists five items to the enquirer and the file holds four. The squadron's own records for 1965 were destroyed on schedule in 1979. A missing letter is a missing letter; it is not evidence of what was in it.
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 TelegraphTHE MAP THAT SAW ANTARCTICA — AND THE AIRMAN WHO SAID SO
Ran the 1965 sentence as an endorsement and built the piece outward from it.
Had that we did not
The quotation itself, transcribed accurately, which is more than most.
Left out
The three hedges in the same paragraph, one of them the word 'apparently'. They are on the same page of the same letter.
-
The Hollow PostThe Squadron Answered Once In 1965 And Has Not Answered Since
Read the silence after September 1965 as a decision taken somewhere above the squadron.
Had that we did not
That the correspondence stops. It does stop.
Left out
The last letter in the file, which is the officer asking not to be quoted without his qualification. Silence after that reads differently.
-
They Buriedthis newspaperPiri Reis Drew A Southern Coast. It Ends Where The Hide Ends.
Paid to have the sheet measured as an object, twice, by people who did not know each other and were not told what answer was wanted.
Had that we did not
Both fits in full, the twelve-band mosaic, and all four 1965 letters transcribed and published.
Left out
We wrote that every reprint since 1966 stops the quotation early. Twenty-one of the twenty-two do. The twenty-second printed the rest of it in 1998 and we found that out from its author, not from our own trace. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Alun Ferreira-Stavely, author of the 1991 study that carries the identification on 3 June 2026 by letter, again on 27 July with the full draft, both fits, the imaging mosaic and all four 1965 letters, and five weeks to answer. Replied 18 August 2026 by email, asking that it run entire. Printed unedited, including the parts about us.
You have measured the thing properly and I am not going to pretend otherwise, so let me concede the measurement first and then tell you what I think you have done wrong.
Sixty-one kilometres against South America and three hundred and forty against the bedrock is not a close call, and two people arriving there separately is worth more than either of them arriving there alone. I accept it. I have accepted it, in print, since my third edition, which is the point I want to make.
You wrote that every reprint since 1966 stops the quotation one sentence early. That sentence was in your draft when you sent it to me and I see it is still in your published piece, softened, in a footnote from your standards editor. In 1998 I printed the whole paragraph — the photograph, the projection, the refusal — as endnote 61 on page 402, and I wrote there that the letter would not bear the weight I had put on it in 1991. Nobody quoted the endnote. Nobody has ever quoted the endnote. You did not find it either; I sent it to you.
That is the thing your desk keeps not seeing. The correction exists. It is printed, it is dated, it is in the same book as the error, and it has no readership at all, because a retraction travels at the speed of the person carrying it and there was nobody carrying mine. You have now built a very expensive apparatus — imaging, two historians, an analyst, an archive trip — to establish something I published twenty-eight years ago at my own expense in the back of a book. I am glad you did. I would rather you had asked me first, and I told you that on the telephone in June, and you did not print it.
On Colonel Ostrowski I want to be plainer, because he has been used badly by everybody including me. He was thirty-four years old, he was asked a courteous question by a stranger, and he wrote a courteous answer with three hedges in it, and one of those hedges is the word 'apparently', which I removed in 1991 for reasons of rhythm. Rhythm. I took a qualification out of a serving officer's sentence because the sentence read better without it, and then eleven other people copied my version, and the man spent the rest of his life being cited for a claim he had explicitly declined to make. He wrote to me in 1979. I did not reply. That is the worst thing in your story and it is mine, not yours, and you were too polite to say so, so I am saying it.
What I dispute is your ending. You say a compiler ran out of parchment. I think you have made the same mistake in the other direction: you have found a mechanism and stopped. Running out of skin explains the bend. It does not explain why a compiler who had eight named source charts in front of him chose to keep going onto a coast he could not fit, rather than stopping at the last name that fitted, which is what a careful man does. Something made him want the extra twelve names on the sheet. Your imaging cannot see wanting. Neither can mine. I would have liked you to say that the animal explains the geometry and not the intention, and I notice that you did not, because it would have cost you your last line.
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.