A signal that filled its telescope’s whole 72-second window exactly once in 1977 — never heard again, by instruments far more sensitive, after the instrument that heard it was torn down
R830’s vaccine at least gets a new facing every year: each formulation meets a season, and a number comes back. Now remove the second facing altogether. What of a test that ran once, on an event that has never recurred, so that its result can neither be inherited (R829), nor re-earned (R830), nor even re-checked — only believed, doubted, or argued about from a single strip of paper? On 15 August 1977, the search for extraterrestrial intelligence produced exactly that, and it has been living on it ever since.
The window was 72 seconds, and the signal filled it. Ohio State’s Big Ear telescope could not track the sky; it sat fixed and let the Earth’s rotation carry a point past its beam, so any source could be observed for just 72 seconds. On that August night a strong narrowband signal did precisely what a distant continuous source should do in such a window: it rose for the first 36 seconds, peaked, and fell for the last 36 as the beam moved away. Days later, volunteer astronomer Jerry Ehman, reviewing the IBM 1130 printout by hand, found the intensity sequence 6EQUJ5 — not a message, despite how often it is misread as one, but a curve: each character is a 10-second average intensity in units of standard deviations above the background, digits up to 9 and letters beyond, so the peak, “U”, means roughly 30 to 31 standard deviations above the noise. The emission sat in a single one of the receiver’s fifty 10 kHz channels, under 10 kHz wide, near 1420 MHz — the hydrogen line, the frequency Cocconi and Morrison had proposed in 1959 as the obvious meeting place for strangers, in a band where terrestrial transmission is forbidden. Ehman circled the sequence in red and wrote one word in the margin: Wow!
Even its address was never quite settled. Big Ear had two feed horns, each seeing a slightly different patch of sky, and a signal should have appeared in both, about three minutes apart. The Wow! signal appeared in one beam and not the other, and the way the data were processed, it can no longer be determined which horn received it. So there are two possible right ascensions for its source, both in Sagittarius, northwest of the globular cluster M55, with the declination fixed at about −26°57′. The one complete test did not even leave a single place to point at in reply.
Every attempt at a second facing has come back empty. Ehman searched for a recurrence with Big Ear itself in the months after the detection and found nothing. Robert Gray searched with the META array in 1987 and 1989, nothing; with the Very Large Array in 1995 and 1996 — an instrument significantly more sensitive than Big Ear — nothing; with the 26-metre dish at Mount Pleasant in Tasmania in 1999, six 14-hour observations, nothing. In May 2022, Breakthrough Listen made its first targeted search, Green Bank and the Allen Telescope Array listening together and apart, and reported no technosignature candidates. This is the shape of the thing: the follow-ups are not weaker reruns of a strong first result. They are stronger instruments returning silence, for nearly fifty years, at both candidate addresses.
And the original test can never be run again at all. Big Ear was dismantled in 1998, the land given over to development, including expansion of a golf course. A different telescope listening at the same coordinates is a neighbour of the 1977 test, not a repetition of it: different beam, different receiver, different radio environment on Earth. Meanwhile the explanations have the same one-shot problem as the signal. The 2017 comet hypothesis was dismissed, including by members of the original team: the cited comets were not in the beam at the right time, comets do not emit strongly at that frequency, and a comet would not explain appearing in one beam only. The strongest current candidate is a proposal, not a verdict: the Arecibo Wow! project reported in a 2024 preprint that Arecibo drift scans in 2020 had found several similar but roughly two-orders-of-magnitude weaker narrowband signals near the hydrogen line, associated with small cold hydrogen clouds, and proposed that a transient flare — from a magnetar or soft gamma repeater — pumping such a cloud could produce a brief maser-like flare in the hydrogen line, possibly the first such event ever recorded. A 2025 reanalysis of the digitised Ohio records sharpened the signal’s parameters and found it stronger than earlier estimates. The project itself is careful about the limit: the work does not close the case and does not conclusively rule out other sources; it reopens the case with a sharper map.
The honest edges. A single occurrence is not nothing. The printout survives, preserved by the Ohio History Connection; the rise-and-fall curve matches Big Ear’s beam pattern, which is why a nearby terrestrial source is hard to make fit, and why Ehman, having first suspected a reflection off space debris, later recanted that scepticism when the requirements on such a reflector proved unrealistic. Nor is this page claiming a transmitter: Ehman himself resisted drawing vast conclusions from what he called half-vast data, and the standing methodological point is blunt — a putative detection that cannot be replicated carries little credibility as a detection, however clean its one appearance was. That is exactly the trap R830’s opening named. The vaccine’s proof expires; the shield’s proof is spent; this proof simply stops. There is a result, it is unusually well documented for a single event, and there is no procedure, at any price in sensitivity, that has yet turned it into a second result. It is believed, or not, on the strength of 72 seconds and a red circle.
The test, schematically. The numbers are schematic units, not measurements: the diagram exists to show what re-observation can and cannot touch. The 1977 record is fixed — no later attempt changes a character of it. Each re-listening, however sensitive, adds one attempt and zero detections. The instrument row ends in 1998 and stays ended.
One detection on the record: 6EQUJ5, 15 August 1977. Attempts to hear it again: 0 so far in this toy — in the world, nearly fifty years of them.
The shield handed its proof to a twin. The vaccine’s makers re-earn theirs every year against a moving opponent. The Wow! signal has no twin, no season, and no opponent who has ever spoken twice. It is the limiting case of the whole sequence: a test the world permitted exactly once, whose entire afterlife is other people listening, with better and better ears, to the place where something was heard once — and reporting, each time, that the place is quiet now.