Here is a haunting that the German state telephone company put on a billing statement. In the autumn of 1967, in a quiet law office in the Bavarian town of Rosenheim, something began running up phone charges that no human hand could explain, switching lights and blowing fuses, and the people who showed up to investigate were not psychics with pendulums. They were physicists from the Max Planck Institute and engineers from the Deutsche Bundespost. The 'ghost,' in other words, was measured on official instruments by people whose entire job is ruling out the impossible.

The location was the law practice of attorney Sigmund Adam. It started with the electrics: fluorescent ceiling tubes that kept going out, fuses tripping, light bulbs swinging. Then it moved to the phones. The office's four telephones would ring at once with no one calling. Calls dropped. And the bills exploded. When the Bundespost investigated, the records showed something absurd: the speaking-clock service had been dialed again and again, in one logged stretch roughly 46 times in fifteen minutes, faster than the office's mechanical rotary phones could physically have dialed it. This is the part that elevates Rosenheim above folklore. The anomaly was not a feeling on the stairs. It was a metered, itemized, billable count.

The Bundespost did what any utility does when the meter says something impossible: it suspected a fault, then fraud. Engineers installed their own call-logging and recording equipment to catch a defective line or a person abusing the phones. Instead their independent gear confirmed the calls were registering even when the office phones were physically incapable of making them, and at one point the recording equipment itself logged events without the dial moving. They eventually ran a dedicated power line into the building to rule out electrical supply problems. The disturbances continued. The state phone monopoly, trying to disprove a haunting, ended up documenting one.

Then came the scientists. Parapsychologist Hans Bender of the University of Freiburg brought in physicists Friedbert Karger and Gerhard Zicha from the Max Planck Institute for Plasma Physics in Munich. They monitored the power supply and witnessed events while the recording equipment ran. Their reported conclusion is the line that has echoed for nearly sixty years: that the deflections and disturbances were real, were not explained by any conventional fault, and that 'some unknown form of energy' appeared to be at work, one that did not behave like ordinary electromagnetism. And Bender noticed the human pattern at the center of it: the phenomena clustered around a 19-year-old secretary, Annemarie Schaberl, intensifying when she was present and following her movements through the building.

Now the skeptical reading, which Rosenheim has never fully escaped. The textbook poltergeist explanation focuses on exactly that detail, the young, unhappy employee at the epicenter, and proposes the most parsimonious cause: a person, consciously or not, generating the effects through ordinary means. A bored or distressed teenager with access to the phones could rack up speaking-clock calls; loose fixtures and old wiring can do dramatic things; and confirmation bias does the rest once a building is declared haunted. Skeptics note that the most spectacular phenomena tended not to occur while the strictest observers were watching the right spot, the classic signature of human agency rather than physics.

But the case resists the easy debunk on one stubborn point: the people who measured it were not credulous. The Bundespost had every commercial incentive to find a mundane cause and clear the bill; it could not. The Max Planck physicists were trained to identify exactly the kind of electrical artifacts a skeptic would invoke; they recorded the events and declined to call them normal. To dismiss Rosenheim entirely, you have to argue that a teenage typist outwitted the German telephone authority's own monitoring equipment and two plasma physicists, simultaneously, for weeks, without ever being caught in the act. That is not impossible. It is just a large claim of its own.

So Rosenheim leaves us with the rarest thing in the paranormal: not testimony but instrumentation. Official meters logged calls that the phones could not have made. Physicists watched deflections they would not certify as conventional. The phenomena stopped when Schaberl left the firm and never recurred there, exactly as a human-centered theory predicts, yet no one ever showed how the human did it under the controls that were running. The unresolved question is precise and it is documented: what, exactly, dialed the speaking clock 46 times in fifteen minutes on a rotary phone that could not dial that fast, and why is it still on the bill?