Sixty-One Second-Hand Accounts. Three People Actually Said Anything.

ROSWELL, N.M. — Sixty-one accounts. Three names at the bottom of all of them.

Between 1980 and 2026, sixty-one people who are the child or grandchild of someone stationed at, or working near, Roswell Army Air Field in July 1947 have given a published statement about what their relative told them. Books, television, documentary, podcast.

We coded every one, for two things: what it claims, and where the speaker says they heard it.

Follow the second column back and the sixty-one collapse into three.

What we counted

Two people on this desk read all sixty-one separately, without conferring, and recorded for each account the claim made and the stated route to it — my father told me, my mother told me what my father told her, I read it and then remembered he had said something like it.

We agreed on fifty-two and disagreed on nine. The nine are published unresolved, with both codings, so a reader can see precisely where the method is soft.

Thirty-eight of the sixty-one trace to Master Sgt. Wilbur Cawdrey-Nunes, a supply sergeant, who died in 1984. Sixteen trace to Ottoline Prewitt-Aguayo, a civilian switchboard operator at the base, who died in 1996. Seven trace to Cpl. Denby Achterberg, who died in 1968 and who never, so far as any record we can find shows, said a word about it into a microphone or onto a page.

The tape

Cawdrey-Nunes is on tape.

In November 1979 a high-school teacher sat in his kitchen with a cassette recorder for thirty-four minutes and asked him about the summer of 1947. The cassette is in a county historical society. We paid to have it digitised and it is published with this story, in full, with a transcript.

He says he loaded crates onto an aircraft. He says a lieutenant told him not to discuss the flight. He says, twice, that he does not know what was in them.

"They took it away," he says, at minute nineteen. "That is what I know. They took it away and I never did find out what it was."

His descendants, in the corpus, report him as having said a good deal more than that.

The symbols

Twenty-two of the sixty-one contain the same specific detail: symbols in purple, running along a beam.

We searched the printed record before 1980 for it — the 1947 newspaper coverage, the 1950s saucer press, three book-length treatments, and the correspondence of two collectors. The detail is not there.

It appears first in a paperback published in October 1980. Every account in our corpus that carries it was given afterwards. The earliest is 1983.

Nineteen ninety-four

Forty-one of the sixty-one were first given in public after 3 July 1994, the night a television dramatisation of the recovery was broadcast to an audience the network put at 24 million.

That is not, by itself, evidence of invention, and we want to be exact about it. Dr. Marisol Threadgill-Obi, who studies what happens to a memory that is questioned repeatedly by interviewers who already hold a view, read our coding sheet and refused to sign up to the headline.

"You have shown that a broadcast produces interviews," she said. "That is not a finding about anybody's mind. It is a finding about television."

What she would say is narrower. A witness asked the same question for forty years by people who want a particular answer will, on average, drift toward that answer without any dishonesty at any point. She calls it the least mysterious result in her field.

What nine people told us

We interviewed nine of the sixty-one at length, between forty minutes and three hours each, and every one of them was easier to talk to than this desk expected.

None of them is lying. That is not a courtesy; it is what the interviews show. What they describe, over and over, is a small sentence about secrecy — we were told not to talk about it — carried by a child, and then asked about, for four decades, by strangers who arrived already knowing what they wanted the sentence to mean.

Marilla Achterberg-Poole, whose father died when she was eleven, put it more precisely than we could. Her father, she says, said one thing and said it once. Every interviewer since has offered her a longer version and asked her to confirm it.

What this desk takes from it

The second generation is not corroboration. Sixty-one accounts standing on three speakers are three accounts with sixty-one people carrying them, and two of those three said noticeably less than they are now reported as saying.

It is not a hoax either, and nobody in this story invented anything. It is an echo, and echoes get louder when you keep shouting into the same canyon and counting the returns.

Rosalind Cawdrey-Nunes Hallett rejects that word, at length, at the foot of this page. Her objection is better argued than our finding and we have not printed a rebuttal to it, because we do not have one.