Words by

Science isn’t a series of lucky accidents

16th November 2026

Let the facts spoil a good dinner party story.

People love accidental discoveries. Penicillin’s discovery by Alexander Fleming is a classic example. Fungal spores wafted in through an open window, fell onto a plate of bacterial colonies, and dissolved them. Another is Louis Pasteur’s discovery of a chicken cholera vaccine, the second vaccine ever developed. After Pasteur left a broth overexposed to oxygen, it went stale. But he ran out of fresh broth, so used it anyway to study the effects of an infection, and found, surprisingly, that it protected the chickens. Or how about the discovery of X-rays? Wilhelm Röntgen was working with a cathode ray tube in a dark room when he suddenly noticed a screen glowing nearby, even though he’d wrapped the tube in cardboard.

Though they’re great stories, I’ve come to believe none of them happened that way. Röntgen had his notes burned after his death, and the story is just a hypothesis. Pasteur was extremely secretive as well, and forbade his family from making his notes and letters public after his death. His grandson, however, did, and after poring over their contents, historians have found that the key experiments were done deliberately by Pasteur’s then-assistant Émile Roux while Pasteur was away for his daughter’s wedding. 

And what about Fleming? His first paper describing penicillin in 1929 never says where the spores came from. Of course, it’s hard to see how he could have known, given they would have been microscopic. But Fleming first told the window story in an interview in 1945, sixteen years after his first report, and his colleagues testified the window was ‘seldom if ever opened’. They weren’t able to replicate the discovery as he described, either: penicillin would have needed to be present on the plate before the bacteria grew, because it blocks bacteria from building their cell walls, rather than killing them once they’ve grown.

Several people have tried to work out what really happened. Fleming’s would-be biographer Ronald Hare made dozens of attempts and found just one reliable way to reproduce the discovery that fit the facts. The spores, he argues, were an unusually productive strain; came from the lab of a fungal researcher on the floor below, rather than the window; settled on the plate roughly when Fleming seeded it with bacteria; Fleming forgot to incubate it; and all this happened during a cold spell in the summer that let the mold grow before the bacteria. Hare noted how much more serendipitous this sequence of events was: ‘the odds against the occurrence of the phenomenon become almost astronomical’, though he ultimately believed it. While writing this column, I read through many of his notes and letters archived at the Wellcome Library and have come to agree with his view: Hare was exceedingly careful in corroborating the events from multiple people and sources from the time.

Part of the reason I find the popular stories fascinating is because it’s jarring to realize how little they’re based on. I do expect some stories of accidental discovery will survive the test of time. In medicine, for example, it’s fairly common to test a drug for one disease and find, unexpectedly, that it works for another: colchicine was used for centuries to treat gout before it was found effective against cardiovascular disease, and sildenafil (Viagra) was being tested for angina when its more famous effect was noticed. Side effects are difficult to predict.

But I wonder why the popular versions have survived so long. Maybe they’re more enjoyable? They’re certainly easier to recount than the other explanations, which are convoluted. A fungal spore floating through an open window to reveal a world-changing antibiotic is a far better dinner party story than a meandering one involving several more coincidences you’ve probably already forgotten. Or the stories make scientists feel just like us: anyone can stumble upon a discovery, and then the challenge is merely to... spend years turning it into something useful?

‘Chance favors only the prepared mind’, as Pasteur put it. I think there’s something to it: the brilliance of these scientists isn’t necessarily in their initial discoveries, but in seeing their potential and doing something with them. Edward Jenner, for example, wasn’t the first to notice that cowpox infections protected against smallpox, but he documented it thoroughly and popularized the method, spreading smallpox vaccination worldwide.

But what bothers me is that calling these discoveries accidents is sometimes the opposite of reality. Pasteur and Roux, for example, carried out hundreds of experiments to find reliable methods to create new vaccines – the popular framing is actively misleading and diminishes their work.

My friend Niko McCarty recently dug into other famously serendipitous discoveries in biology, such as those of green fluorescent protein and nanopore sequencing, and explained they came from systematic trial and error. He also describes the idea of ‘engineered serendipity’: in optogenetics, for example, ‘the key breakthrough came about because the inventors wrote down the exact requirements they were looking for [...] and then enumerated, or wrote down, all the ways they could possibly imagine to make this happen’. It makes me wonder if learning how they really happened could help understand how to make more discoveries systematically.

Yet I’m still left with questions. Why credit chance events instead of hard work? Why hide or burn your notebooks?

Messiness, mistakes, and privacy come to mind. But here’s another theory: competition and rivalry. Gerald Geison, a historian of Pasteur, suggests the method he ultimately used to develop chicken cholera vaccines was ‘remarkably simple’ and could be applied to other diseases too: ‘had he revealed immediately the method of attenuation through exposure to ordinary air, Pasteur might have faced a host of competitors in the search for other vaccines’. Pasteur was probably the best-paid scientist in France at the time, and sold his inventions commercially. His secrecy might have been rational: patents weren’t allowed on medicines in France then, so keeping the method secret would have helped him maintain his monopoly on producing them. Pasteur was so cagey that he didn’t allow his collaborators to take notes, took his notebooks with him when he traveled, and published his methods (without much detail) only months after announcing his discoveries.

Meanwhile, Wilhelm Röntgen, who discovered X-rays and had his notes burned after his death, had been in a long dispute with another scientist, Philipp Lenard. Lenard had spent years researching cathode rays before him, and shared his methods and discharge tube with Röntgen. Once Röntgen became famous, Lenard accused him of having robbed him of the discovery. That dispute, historians suggest, could have motivated the note-burning, but it’s hard to know without being able to read them. (Incidentally, Émile Roux, too, had his notes burned periodically, and it’s only because Pasteur’s were released that we know what happened at all.)

Science seems like it should be above the messiness. It’s an institution we’ve built to guard against our biases. But it’s also just people – who forget, hide their notebooks from their colleagues, burn their papers, and spend years fighting over who deserves credit. It’s shaped by rivalries, incentives, obsessions, and years of unglamorous work, most of which is forgotten. That, to me, makes it far more interesting than a story about an open window.

Get the print magazine

Subscribe for $100 to receive six beautiful issues per year.

Subscribe