The Ship Was Being Demagnetised. The Three Annotators Were Probably One Man.
Thirty-one copies of a 1957 mimeograph survive. A document examiner, told nothing about the book, compared 640 characters in three inks and found one hand at three speeds. The vessel was at a deperming pier.
A months-long reporting project. Documents cited below are held in The Vault and available to readers.
GARLAND, Texas — Thirty-one copies survive of a run of 127, mimeographed in 1957 inside a defence electronics firm in Garland, Texas, for circulation in its own building.
This newspaper photographed the margins of all thirty-one, both sides of every annotated leaf, at 1,200 dots per inch.
Every published account of the Philadelphia Experiment descends from those margins. Not most. All of them.
The three hands
The object underneath the mimeograph is a 1955 paperback by an astronomer named Ellery M. Sowerby, remaindered within a year of publication, annotated by hand in three inks — blue-black, green, and a red that has gone brown.
Roughly four thousand one hundred words of marginalia. The 1957 preface, written by whoever prepared the run, labels the writers A, B and C, treats them as three readers who have had the book in turn, and notes that they appear to disagree.
They do disagree. That is the striking thing about the pages and it is why they have been reprinted for seventy years.
What the examiner was given
We sent a forensic document examiner three sets of exemplars, sorted by ink and labelled only by ink. No title page. No context. No indication that anything about the material was contested.
Annelise Storrow-Baptiste has done this work for twenty-nine years and has been qualified as an expert in eleven jurisdictions. She compared 640 characters across the three sets, on letter formation, pen lift, connecting stroke and the position of the pen in the hand.
She reported one writer, at three writing speeds, using three pens.
Her opinion is probable — level four on the nine-point scale her profession uses, one step short of identification. Pressed for a number she said about eight in ten, and asked that the number appear in print.
The eleven letters
Sowerby's papers are in a university manuscripts collection in Michigan, in boxes nobody had called up since 2004.
Boxes 9 to 11 hold eleven signed letters from one correspondent between 1955 and 1957: a merchant seaman named Vance Orlin Prewitt, writing from a rooming house in Chester, Pennsylvania. Sowerby's carbon replies to the first six are filed beside them.
We photographed the letters and sent them to Storrow-Baptiste blind, mixed into a set of nine other period hands. She matched them to writers A, B and C without hesitation and without knowing what she had matched.
What the ship was doing
The escort vessel at the centre of the account was not missing on the disputed dates in October 1943. She was moored at the yard's deperming slip, being ranged.
Her magnetic treatment record card survives, with the readings before and after. So does her deck log, entered every four hours in the officer of the deck's hand. Both are reproduced with this story.
Deperming is the deliberate scrambling of a steel hull's permanent magnetism by wrapping it in cable and putting current through it, so that a magnetic mine sitting on the bottom does not notice a ship going over the top. It was routine, it was scheduled, it generated paperwork, and it worked.
The yard's own house newsletter for March 1944 runs a photograph of the slip under the heading Making Them Invisible.
What actually happened here
A ship disappeared from magnetic mines. A man in a rented room read a paperback about it and wrote in the margins at different hours, in whatever pen was nearest, and posted his questions to an astronomer who answered him kindly for two years and then stopped.
A firm ran off 127 copies for its own people. Seven books quoted them. A novel came out of the books and a film came out of the novel, and the film's dialogue is writer B nearly word for word.
The chain is shorter than the chain we traced for the Lovelock skeletons in PX-2011, and it is the same shape: one physical object, then seventy years of people quoting each other.
The draft we sent to Prewitt's niece used the word hoax once. She told us to stand behind it or take it out, and we have taken it out, which is why the ninth paragraph her letter answers is no longer in this piece. A documented hoax looks like PX-2089: two people, a plan, a target, and a letter afterwards in which they congratulate each other. There is nothing of that shape here.
What this desk takes from it
Our headline carries the word probably because Storrow-Baptiste asked for it in writing, and because she is right that the distance between probable and identification is her entire profession.
It does not carry her number. She gave us about eight in ten and asked for that in print too, and a headline is not where a number of that shape survives. It is here instead: about two findings in every ten graded this way turn out to be wrong.
Prewitt's niece has read all of this and her answer is below. She does not dispute the handwriting. She disputes what we think the interesting question is, and on that she may well be right.
Sowerby stopped writing in the spring of 1957. Prewitt sent four more letters into the silence. The fourth runs to nine pages, and it is in the library, and anyone may call up box 11 and read it.
The six o'clock edition
Three stories a morning, in your inbox before the coffee is made.
Sources & Method
We did not read the margins for what they say. We photographed them — all thirty-one copies, both sides of every annotated leaf, at 1,200 dots per inch — and gave a document examiner three sets of exemplars labelled only by ink colour, with no title page, no context and no indication that the material was contested. She was told what the book was after she had filed. We then took the vessel's own paperwork for the disputed week out of the archives and read it against the story it is supposed to hide.
Who we spoke to
- Annelise Storrow-Baptiste, Forensic document examiner, twenty-nine years, court-qualified in eleven jurisdictions. Commissioned and paid in advance; given unlabelled exemplars with no context and not told what the book was until after she reported February–May 2026 Reported at level four on the nine-point opinion scale — probable. Pressed for a number, she said about eight in ten and asked that the number be printed.
- The Sowerby papers, boxes 9 to 11, Manuscripts division of a university library in Michigan. Read in the reading room over six days; the eleven signed letters photographed under the collection's standard reproduction licence April 2026 Eleven letters from one correspondent, 1955 to 1957, all from the same rooming-house address in Chester, Pennsylvania. Sowerby's carbon replies to the first six are filed with them.
- Magnetic treatment record card and deck log, October 1943, Ship's records for the escort vessel at the centre of the account. Ordered as certified copies from the national archives, $214 in fees; both reproduced in full with this story March 2026 The yard's own house newsletter for March 1944 runs the deperming slip under the heading 'Making Them Invisible'. The phrase is the yard's, not the legend's.
- Rowan Devereux-Ito, Antiquarian bookseller; has kept a census of the 1957 run since 1994. Interviewed three times and gave us the locations of the copies; took no fee and asked for none January–June 2026 Locates 31 copies in eight collections. Believes at least forty more exist in private hands and says so every time we quote him on thirty-one.
- Hazel Prewitt-Rennard, Niece of Vance Orlin Prewitt, merchant seaman, 1918–1988. Interviewed at her house in Wilmington, Delaware, and given the full draft, the examiner's report and photographs of all eleven letters June and July 2026 Holds his side of the correspondence — four of Sowerby's replies, kept in a shoebox. Her answer is printed at the foot of this page.
Documents
- PX-2011 — Chain of printed sources for the Lovelock 'giants', 1911–2024 accepted
- PX-2089 — Letter from a small-press publisher to a young collaborator, on the reception of a hoax the two of them had placed in print accepted
What we could not confirm
- Ninety-six copies of the run are unlocated. Thirty-one is a sample and we have described it as one throughout, but it is not a census, and a single copy carrying a fourth hand — or carrying the three hands in a different distribution — would undo the central finding of this story.
- One hand, and not certainly. Storrow-Baptiste reported probable — level four on a nine-point scale, one step short of identification — and, pressed, about eight in ten. The distance between probable and identification is the entire professional structure of her discipline, and about eight in ten means that of every ten findings she grades this way she expects to be wrong about two.
- We cannot date the marginalia. The mimeograph is 1957 and the paperback is 1955, and everything in between is inference. Three writing speeds is a finding about pen pressure and letter formation. It is not a finding about three different nights, and every sentence in this piece that implies otherwise is ours, not hers.
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 TelegraphTHREE MEN SAW IT — AND ONLY ONE OF THEM WAS EVER FOUND
Treated the identification of one annotator as the disappearance of the other two, and asked what became of them.
Had that we did not
A clear and honest account of the mimeograph's physical history, better than ours on the printing.
Left out
That the examiner's report, which they had, says one writer, and says it on the second page.
-
ApertureThree Inks, One Room
An essay on the marginalium as a form, and on writing to a scientist who has stopped writing back.
Had that we did not
The best paragraph anybody has written about the green ink, which is the ink he used when he was calm.
Left out
The deck log, the treatment record, and the entirely mundane thing the ship was doing that week.
-
They Buriedthis newspaperThe Ship Was Being Demagnetised. The Three Annotators Were Probably One Man.
Photographed thirty-one copies, gave the hands to an examiner blind, and then went and got the vessel's paperwork.
Had that we did not
The examiner's full report, including her stated confidence and the second condition we did not meet.
Left out
Storrow-Baptiste asked for two things in the headline and got one. Probably is there. About eight in ten is not, and a reader who stops at the headline has no way of knowing how wide eight in ten is. The number was the thing she asked for first. — V. Ashcombe-Doyle, standards editor
Right of Reply
They Buried contacted Hazel Prewitt-Rennard, niece of Vance Orlin Prewitt on 11 June 2026 in person, again on 24 July with the full draft, the examiner's report, photographs of all eleven letters and four weeks to answer. Replied by letter on 14 August 2026 and asked that it run without cuts. Printed unedited.
You have written eight hundred words about handwriting and you have not once used the word lonely, so I will use it for you.
My uncle went to sea at seventeen and came off the ships at thirty-four with a bad hip and no trade. He rented a room in Chester with a hotplate and a window onto a wall. He read the book because it was on a spinner rack at the drugstore for thirty-five cents and it was about something enormous, and he wrote to the man who had written it, and the man wrote back. That is the whole event. A stranger with a doctorate answered a merchant seaman's letter within the fortnight, and then answered the next one, and then the next, for two years.
Your examiner is right and I am not going to argue with her. Three inks, one hand. I have his shopping lists. I know that hand.
But you have arranged your story so that the three inks are the crime and the ship is the punchline, and the man in the middle of it is a mechanism for producing marginalia. He was not pretending to be three people to deceive a defence contractor who had not yet heard of him and would not for two years. He was reading a book at two in the morning and answering himself in the margin because there was no one else in the room to answer him. Anybody who has lived alone knows what that looks like on paper.
I want to say something about the word hoax, which you use once, in your ninth paragraph, and then drop. You should either stand behind it or take it out. A hoax has a mark and a payoff. Show me either. He gave the book away. He never sold a copy, he never sat for an interview, he refused two, and when the film came out in 1984 he was living in a rooming house in Wilmington nine blocks from where I am writing this and he did not tell his own family that any of it had anything to do with him. I found out in 1991, from a box.
And on Dr. Sowerby, since you have his side of it in a library and I have four of his letters in a shoebox. He was kind. He was busy and he was kind and his handwriting was terrible. Then the letters stopped, in the spring of 1957, and my uncle wrote four more into the silence, and the fourth one is nine pages, and I would rather you had not read it, and I understand why you did.
He stopped answering because he was ill and then because he was dead, and my uncle did not learn that for nine months, and when he did he wrote it on the flyleaf of his own copy with the date, in the green ink, which your examiner will tell you is the one he used when his hand was steady.
Print the deck log. Print the pier. I have no quarrel with the ship. Just do not tell me the interesting question is how many people were holding the 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.