Hundreds of people wrote to a Bristol newspaper in the 1970s about a hum they could feel more than hear — a warehouse fan was blamed, the warehouse closed, the reports continued, and fifty years later the same complaint has been filed across continents by people who have never met

Hundreds of people in Bristol spent the mid 1970s writing letters about a noise they could feel in their chest and not quite hear.

Those letters went to the Bristol Evening Post, and they described much the same thing: a low mechanical drone, worse at night, worse indoors, and completely inaudible to whoever was asleep in the next room. Experts drafted in at the time blamed factory noise, electricity pylons and tinnitus. Less cautious voices suggested flying saucers parked over the city, as the BBC later wrote in a look back at the affair. Bristol City Council sent environmental health officers out with meters in April 1980 after another wave of complaints, and the same broadcaster has since reported that residents in Staple Hill and Downend were still hearing it decades later.

Large industrial fans at a nearby warehouse were the early favourite, a theory recounted by The Week. It never fully explained the phenomenon.

What happened in Kokomo

Kokomo, Indiana, got its own version starting in 1999, and the city did the sensible thing. It paid an acoustics firm to go looking. Consultant James Cowan named two culprits: a 36 hertz tone from cooling tower fans at the local DaimlerChrysler casting plant, and a 10 hertz tone from air compressor intakes at Haynes International. Both companies adjusted their equipment.

Complaints kept arriving anyway, as Atlas Obscura has documented. It is the same pattern that dogged the original warehouse theory in Bristol: a plausible mechanical culprit, and a mystery that did not fully go away.

What Taos measured

The most rigorous attempt came from New Mexico, after residents of Taos petitioned Congress in 1993. A team led by engineering professor Joe Mullins of the University of New Mexico, joined by staff from Sandia and Los Alamos, interviewed hearers and blanketed the town with microphones, geophones and magnetometers.

Around two per cent of residents reported hearing it, and many could reproduce their private tone on a signal generator with unsettling consistency. Nothing in the environment matched. In a 2008 conference paper on noise as a public health problem, Cowan summarised the outcome: no acoustic signals corresponding to what people described were detected by any instrument, and no source was ever identified.

The one case that closed

Windsor, Ontario, is the exception that keeps everyone honest. Complaints began around 2011, describing a pulsing rumble that shook windows on the city’s west side. Canada’s foreign affairs department funded a study by the University of Windsor and Western University, and the study’s published summary concluded that the hum existed, and that its likely source was blast furnace operations on Zug Island, a heavily guarded industrial island in the Detroit River.

Then in 2020 U.S. Steel idled the plant. Lead researcher Colin Novak told CBC News that the noise practically vanished.

Sometimes it really is a machine on the far side of a river, running hot.

The case that stays open

Everywhere else, attention has drifted from the environment to the ear. In March 2026, a team including Bonifaz Baumann and Markus Drexl published a study in PLOS One that took 28 German volunteers who hear the noise and tested two mechanical explanations. Perhaps their low-frequency hearing was freakishly good. Perhaps they were hearing spontaneous otoacoustic emissions, the faint sounds a healthy inner ear produces as a by-product of amplifying everything else.

Both came back empty. Only two participants showed unusually sharp low-frequency sensitivity, and no low-frequency emissions turned up in anyone. What the volunteers had in common was a median matched pitch of about 50 hertz and no measurable cause, which pushed Drexl’s group toward subjective tinnitus in the bass register: a phantom sound that happens to impersonate machinery rather than ring like a struck glass.

One study of 28 people in one country is not a verdict, and the authors know it. They have not ruled out genuine external sources, and Windsor is proof that those exist.

Scale is the other problem. Glen MacPherson, a Canadian schoolteacher, has spent more than a decade building the World Hum Map and Database Project, logging thousands of self-reported entries from places as scattered as Iceland, New Zealand, Kazakhstan and the Philippines, per Global News. He is the first to point out the data skews English-speaking and self-selected. It is a map of who filled in a form, not a census of a sound.

What Drexl keeps returning to is the gap underneath all of it. Almost everything audiology knows about hearing was built on higher frequencies, on speech and music and the ringing kind of tinnitus. How the ear handles sound below 250 hertz, and infrasound below 20, remains thinly charted. Fifty years of letters from five continents may end up mattering less as a mystery than as a complaint filed against a blind spot, one that hearing science never got around to closing.