The Elephant Who Took 3,000 Doses
On August 3, 1962, researchers Louis Jolyon West and Chester Pierce dart-fired 297 milligrams of LSD into Tusko, a 14-year-old, roughly 7,055-pound (3,200 kg) Indian elephant at Oklahoma City Zoo. They were testing whether LSD could trigger musth, the violent state male elephants periodically enter, and scaled the dose up from a human recreational amount using body weight alone.
Five minutes later, Tusko trumpeted, collapsed, and never rose again, dying about ninety minutes later despite 2,800 milligrams of an antipsychotic and a barbiturate meant to revive him. It remains the largest single LSD dose ever given to a living creature.
Oddly, other elephants later given LSD under more careful conditions survived without incident. Same species, same drug, so what made Tusko’s case different?
Why Scaling by Weight Isn’t Enough
The logic looked sound: a heavier animal should tolerate proportionally more of a drug. But body weight and drug sensitivity don’t scale together predictably across species, liver and bloodstream handling depends on biology, not size. Elephants may simply be unusually sensitive to LSD for reasons unrelated to mass.
Whether the LSD itself or the follow-up drug cocktail killed Tusko remains unsettled. Either way, the musth hypothesis wasn’t the failure, the math converting a human dose into an elephant’s body was.
The Monster Study: Orphans as Lab Rats
In 1939, psychologist Wendell Johnson and graduate student Mary Tudor experimented on 22 children at the Iowa Soldiers’ Orphans’ Home without consent. Ten already stuttered; of the 12 fluent speakers, six were told repeatedly they were starting to stutter and must stop, while six served as controls.
Johnson believed stuttering could be triggered by judgment rather than physical flaw. By that measure the study failed, none developed an actual stutter, but several fluent children came away with lasting anxiety and speech problems.
Never published, the study stayed hidden until a 2001 San Jose Mercury News investigation exposed it. By August 2007, six survivors settled with Iowa for $925,000, with the University of Iowa issuing a public apology. The failed hypothesis isn’t the scandal, what happened to the children is.
The Chimp Who Humanized the Baby
In June 1931, psychologists Winthrop and Luella Kellogg raised a seven-and-a-half-month-old chimpanzee, Gua, alongside their ten-month-old son Donald, giving both identical baths, meals, lessons, and tests, to see if a human environment could make a chimp act human.
Over nine months, Gua even outpaced Donald on some tests but never learned to speak. The real surprise was Donald, who began grunting and adopting Gua’s habits instead.
The Kelloggs ended the experiment in March 1932 as Donald’s chimp-like behavior grew hard to ignore. The hypothesis assumed influence would run only one way, from household to chimp. With just two children, there was no way to separate “environment shapes behavior” from “two toddlers raised together copy each other.”
The Doctor Who Drank Vomit
In the early 1800s, Dr. Stubbins Ffirth tried to prove yellow fever wasn’t contagious, reasoning that its seasonal fade didn’t match typical contagious diseases. He exposed himself to infected patients’ vomit, blood, urine, saliva, and sweat, ingesting, injecting, and inhaling it, and never got sick, concluding the disease wasn’t contagious.
The real explanation came later: his samples came from patients past the contagious stage. Yellow fever actually spreads via mosquitoes carrying a virus, undetectable with tools of his era. Ffirth got a clean, repeatable, but wrong result, not because contagion wasn’t real, but because he sampled the wrong moment in the illness.
The Baby Box That Worked, Until the Headline
In the 1940s, B.F. Skinner built an enclosed, climate-controlled crib with a safety-glass front for his daughter Deborah, who spent roughly her first two years in it. She grew up healthy, later marrying a professor and becoming an artist in England.
The device wasn’t the problem, a 1945 Ladies’ Home Journal article, “Baby in a Box,” combined with Skinner calling it an “experiment,” made it sound like a lab cage. Backlash followed, along with a false, persistent rumor that Deborah later died by suicide. She hadn’t.
Cribs went into commercial production anyway, with over 300 children raised in them at about $150 each. A later Psychology Today survey of 50 such families found satisfied parents. None of that undid the damage from one headline, the science was sound, the messaging sank it.
What These Five Have in Common
Tusko died from a dosing error. The Monster Study failed on ethics, not evidence. Gua and Donald exposed an untested assumption about who influences whom. Ffirth’s result was a false negative from testing the wrong moment in an illness. Skinner lost a sound invention to one headline. In every case, the idea under test wasn’t the weak link.
It’s easy to blur “the hypothesis was wrong” with “something around the hypothesis went wrong.” Asking what specifically broke, the dose, the ethics, the sample, the spin, gives a sharper read on how research actually fails.
So the next time a famous experiment gets dismissed as proof some idea was ridiculous, it’s worth asking which of those things really broke. Usually, it isn’t the idea.