The Radio Echo That Arrives Three Seconds Too Late

What We Know

Jørgen Hals was a civil engineer in Oslo with a shortwave receiver and a habit of logging what he heard. In late summer of 1927, tuned to the Dutch station PCJJ in Eindhoven, he caught the ordinary echo any shortwave listener of the era would expect: a repeat of the signal about a seventh of a second later, the wave’s own trip around the curve of the Earth. Then, about three seconds after that, he heard the signal again.

Three seconds is a long time for something moving at the speed of light. In that interval, a radio wave could circle the Earth roughly twenty-two times. Hals wrote to Carl Størmer, a Norwegian physicist already known for his work on the aurora, and told him what he’d heard. Størmer took it seriously enough to publish it in Nature the following year, that November, and the report reached Balthasar van der Pol, a physicist at Philips in the Netherlands, who joined the search. Between 1928 and 1930 the two ran what one later physicist called probably the largest coordinated effort ever mounted to chase the effect: scheduled transmissions, timed and logged simultaneously by listeners in two countries. On the night of October 24, 1928, it worked. Stations in Norway and the Netherlands recorded the same delayed echo, on the same signal, at the same time.

That correlation is still the single best piece of evidence anyone has produced. It didn’t explain anything. Van der Pol couldn’t pin down a cause, because the delays refused to repeat on demand, but the correlation proved two things: the reports weren’t one man’s faulty equipment, and whatever caused them could be caught by more than one listener at once, if the timing lined up.

It’s happened again, sporadically, ever since. Licensed amateurs on multiple continents have logged the same signature: an ordinary echo, then a second one, seconds later, with no obvious trigger, across nearly a century of radio history. What none of them, working alone or together, has managed is to make it happen twice.

The Roadblock

Radio physics has a real answer for short echoes: a signal can travel along Earth’s own magnetic field lines and reappear at the matching point in the opposite hemisphere, arriving a fraction of a second late. That mechanism is well understood and independently confirmed, the best-characterized candidate in the entire literature. It also tops out around half a second per trip. Reaching Hals’s three seconds means stacking that mechanism many times over, and each hop should bleed off signal strength; no one has shown how a shortwave transmission survives that many hops and still arrives loud enough to hear.

The harder problem is that the echo won’t sit still to be studied. It doesn’t arrive on a schedule, doesn’t repeat on request, and by the time a listener recognizes what they’re hearing, it’s usually over. The single strongest data point in the record, the October 1928 dual-country reception, has never been repeated under conditions rigorous enough to satisfy a physicist starting from scratch. A dedicated hunt to catch and instrumentally verify one, published in 1985, found that almost every individual case, reviewed closely enough, turned out to have a more ordinary explanation available. That doesn’t prove nothing real is left over. It proves that hearing it is far easier than proving it.

Best Guesses

No investigator, amateur or professional, has ever caught a long delayed echo on command. What exists instead is a century of reports clustered tightly enough to look like one phenomenon: named observers, similar delay windows, no repeatable trigger. Four explanations have survived contact with that record. None has ever been confirmed, and none has been fully ruled out.

Distant plasma cloud reflection

Størmer’s own hypothesis, refined through van der Pol’s 1928–30 campaign: somewhere beyond the ionosphere, a temporary region of ionized gas, proposed at distances up to and past the Moon’s orbit, catches the signal and sends it home, the extra distance accounting for the extra seconds. No instrument has ever independently detected a cloud with the density or persistence the theory needs — it’s inferred backward from the delay, never observed directly. What the theory does have is the best-documented data point in the whole case: it’s what drove two working physicists to organize the simultaneous Norway–Netherlands reception of October 1928, still the single most tightly corroborated echo on record.

Multi-hop magnetospheric ducting

The one confirmed mechanism on this ballot, just run further than anyone has directly measured. Magnetospheric ducting, in which a signal is guided along Earth’s own magnetic field lines to a matching point in the opposite hemisphere, reliably produces echoes delayed by a fraction of a second, and it’s the best-understood candidate available. Stack enough of those hops end to end and the delay adds up to Hals’s three seconds or more; amateur radio operator Alan Goodacre’s 1980 observations logged a pattern consistent with dozens of circuits. The catch: each hop should bleed off signal strength, and nobody has shown how a transmission survives that many hops and still arrives audible.

Equipment or multipath artifact

The skeptical case, and the best-tested one: there is no exotic phenomenon here, only ordinary ionospheric multipath, receiver ringing, or a logging error, misidentified as one clean delayed echo. It doesn’t explain why competent, independent listeners across nearly a century keep reporting delays that cluster around the same few seconds, rather than the scattershot errors bad equipment should produce. What it has instead is the record’s most rigorous test: in a 1985 paper, physicists Robert Vidmar and Frederick Crawford described a dedicated Alaska campaign built to catch and verify one, and found that almost every individually reviewed recording had a mundane cause available.

Real but uncaught rare event

Maybe nothing here is wrong except the sample size. The reports span a century and multiple continents, including a case investigated within months by two physicists who still couldn’t force it to repeat on demand. Modern always-on digital monitoring has existed for only a sliver of that reporting window. On this view, the effect is real and physical, just rare enough that a fully instrumented, corroborated capture hasn’t happened — including now. The trouble is that this claim can’t be disproven by any number of null results, which makes it satisfying to hold and hard to confirm.

The Verdict

A single receiver, however careful its owner, can always be second-guessed: ringing, an ordinary echo, a clock that drifted. What can’t be second-guessed is several independent, GPS-synchronized digital receivers, in different places, timestamping the same transmission and its delayed return, with the originating signal logged from source to receiver. That is the exact setup the HamSCI amateur-radio research network is trying to build, nearly a century after Hals wrote his first letter to Størmer.

Until a capture like that exists, every case in the record, October 1928 included, keeps its asterisk: heard, logged, corroborated by a second listener when the physicists got lucky, and never once caught by an instrument that couldn’t be argued with. The record still rests on people who heard it. No instrument ever has.