In my opinion, we're possessed by a cultural epidemic of think pieces doing rich and nuanced science history, but wrongly framed in the form of correcting "myths" that, in their substance amount to quibblings over narrative emphasis. It's easy to get taken in by the framing because it truly is enlightening, and the argument goes down so smooth because its embedded in a rich, curious, and fascinating scientific history that otherwise embodies best practices I would happily celebrate.
But the key details about the story of penicillin are that a moldy plate showed bacteria-free clearing, Fleming saw it, isolated the mold, proved its germ-killing filtrate and published the finding, which is the heart of the story and which is not a myth.
I'm sure it's true enough that St Mary's windows were usually kept shut to keep pathogens in and contaminants out, that London's August 1928 cold snap would have slowed staph growth, that Fleming's first notes Or 8 weeks later than the actual event, and that a modern plate seeded with bacteria first will not produce the celebrated halo unless the mold is given a head start. The article makes much of the fact that today’s researchers cannot reproduce the famous halo if they add staph first, yet that difficulty rebuts a sequence Fleming never claimed to have used.
These points are significant, even fascinating, yet the article inflates them into a strobe-lit "MYTH" banner, turning normal human imprecision about times and temperatures into evidence of wholesale fiction, which abuses the ordinary friction of any retrospective account and punishes the story for the very human messiness that makes it instructive.
The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
It's a myth in the most literal way. Fleming published and promoted his results despite a lack of reproducibility. By the time he won the Nobel Prize, he had backformed or misremembered a folksy story about an open window. That's textbook mythmaking.
It can both be fine to have a glib story to tell schoolkids and important to recognize that the actual intellectual process is messier and more complex.
I have now actually read Flemming's 1929 manuscript that first described penicillin [0]. It is a careful and well documented scientific report describing the action of penicillin on various species of bacteria, how to produce it, and some of its chemical properties. It describes how penicillin can kill bacteria isolated from the throats of nurses, and shows that it has low toxicity in mice, and is possibly safe for use in humans: "Constant irrigation of large infected surfaces in man was not accompanied by any toxic symptoms, while irrigation of the human conjunctiva every hour for a day had no irritant effect."
It is far from having a "lack of reproducibility" and in fact allowed others to quickly and accurately replicate his discovery.
The path to his discovery may have been difficult to replicate, but the fact that the mold could kill other bacteria was not, and was immediately replicated.
It just wasn't seen as relevant because, at the time, few people imagined its internal use in humans and it was instead seen more as a tool for other microbiologists and the like. The jump to "And then I see the disinfectant, where it knocks it out in a minute. One minute. And is there a way we can do something like that, by injection inside or almost a cleaning?" took quite some time.
> The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
This is the same conclusion as the article. IMO, the importance of challenging the myth is that it has hisotrically taken precendence over your (and the article's) conclusion.
FTA
> Fleming’s 1929 penicillin paper may have been written as a linear process, but that’s almost certainly not how the discovery occurred. And by eliminating these complicated twists and turns, Fleming inadvertently obscured what may be one of the most important lessons in scientific history: how combining a meticulous research program with the openness to branch out into new directions led him to Nobel Prize-winning success. Neither rigid plans nor the winds of chance are enough on their own; discovery requires both.
> The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
Personally for me, while less important, I really appreciate the investigation into the narrative.
I agree that the science is more important and the results don't care about the story.
The balance is that we don't need to go around correcting everyone, but knowing more about the details of the story is worth my time in reading this piece. I think the article strikes the right tone.
To be anal about being anal, the article doesn't preclude Fleming's account. It argues that it's unlikely, but countless highly improbable things are happening every second. On this topic somehow Ancient Egyptian poultices (and in cultures onward - though they are the oldest recorded account) even used moldy bread to treat bacterial infections, somehow stumbling onto genuine antibacterial aspects for an absurdly counter-intuitive treatment that has a real effect. However it was initially discovered back then, let alone replicated and confirmed, must have been through an unimaginably improbable series of events. Yet it happened. That's rather the story of humanity.
Got a quick insight about how penicillin works: interferes with cell-wall building which is a destroy and recreate process by preventing the recreate part.
Got a quick view into the scientific process and communication: Fleming focused on the insight - penicillium kills staphylococcus - and left out the circuitous detail. This is important so that the big win here is very clear.
And got an insight into human nature and memory: Fleming didn’t tell the accidental contamination story until much later. It could possibly be even an idea someone else might have come up with which then took root in his mind (ironic haha!)
The communication aspect reminds me of Mendel’s far too perfect ratios for his pea plants. That kind of “repeat till difference clear” statistics would be decried today but perhaps that was to communicate rather than to determine.
And finally, I really enjoy reading about human process innovation because I think it’s a big factor in how Humanity grows. The lab notebook has to be some kind of star performer here - Fleming’s notes allow us to look back like this.
When I experiment with things, I naturally lean to keeping notes on my test protocol, observations, and results. But not because of some personal genius. It’s just the standard way I was taught as a child in our science labs.
I won’t claim to the rigor of a microbiology lab but even just the process notes help a lot, which is useful since I’m just testing molecules on myself.
If you are not familiar with more of Mendel or plant biology, he got extremely lucky in picking a two chromosome species. The next plant he picked had more than two chromosome types so he spent the rest of his life hitting his head against the wall - obvious to us but him not having a theory and expertise with microscopes to explain his pea results hampered him greatly beyond his initial pea plant studies.
What do you mean by a two chromosome species? A quick google says pea plants have 14 chromosomes. I only looked because I had never heard of a species only having two chromosomes. Do you mean the traits he was selecting for only had two alleles?
The traits he picked- hope I get this detail pedantically correct: had only two alleles, each allele had an obvious phenotype controlled mostly by that gene, the two phenotypes were binary (no intermediate "half-wrinkled-half-smooth"). and all segregated independently (different chromosomes, or far enough that the linkage was extremely weak). I remember my high school teacher speculating he inspected many different phenotypes and then reported his results on the final ones he picked where the results were nice.
Unfortunately, most modern genotypes and phenotypes in humans don't follow these patterns, and over the years, genetics devloped an entire vocabulary and physical model to explain them, although at a fairly abstract level. None of it made any sense to me so I follow the biophysics/molecular biology approach which tends to consider many more underlying physical details
Diploid. Peas have 7 pairs of chromosomes. Many types of plants have multiple (polyploid) copies of each chromosome, and you will not discover the classic Mendelian dominant-recessive pattern by studying them. Mendel lucked out by studying a diploid species. https://duckduckgo.com/?q=polyploid&ia=web
Right, my bad for wrong terminology as biology nor botany is not my speciality either. I was thinking along the lines of XY for humans, then the ordinary two row, two column chart used to teach the basics of Mendel with pea plants and dominant/recessive non polygenic traits in introductory biology classes.
It's a shame that Fleming misremembered his process of discovery and created a myth of accidental discovery.
I like the Root-Bernstein narrative more. That in the monotonous execution of routine experiments for something unrelated an unusual observation 'forced' them to discover penicillins antibacterial properties.
Not an accidental discovery by good fortune in a serendipitous sense. An accidental discovery of a brute force exhaustive search. The narrative of we spent months meticulously examining hundreds of samples is less romantic, but is one that supports the importance of funding scientific inquiry.
We won't make progress by hoping people leave culture plates out on window sills. We make progress when we fund meticulous exhaustive efforts of discovery.
> Mendel’s far too perfect ratios for his pea plants.
"Remember, if you flip a coin 200 times and it comes heads up exactly 100 times, the chances are the coin is actually unfair. You should expect to see something like 93 or 107 instead".
Isn't 100 / 100 the most likely outcome for a fair coin? Sure it's unlikely that you hit exactly that result, but every single other result is even less likely.
What I'm trying to say: if you get 100 / 100, that's not a sign of an unfair coin, it's the strongest sign for a fair coin you can get.
Okay, but it doesn't make sense to arbitrarily group together some results and compare the probability of getting any 1 result in that group to getting 1 particular result outside of that group.
You could just as easily say "you should be suspicious if you flip a coin 200 times and get exactly 93 heads, because it's far more likely to get between 99 and 187 heads."
It's suspicious when it lands on something that people might be biased towards.
For example, you take the top five cards, and you get a royal flush of diamonds in ascending order. In theory, this sequence is no more or less probable than any other sequence being taken from a randomly shuffled deck. But given that this sequence has special significance to people, there's a very good reason to think that this indicates that the deck is not randomly shuffled.
In theory terms, you can't just look at the probability of getting this result from a fair coin (or deck or whatever). You have to look at that probability, and the probability that the coin (deck etc.) is biased, and that a biased coin would produce the outcome you got.
If you flip a coin that feels and appears perfectly ordinary and you get exactly 100 heads and 100 tails, you should still be pretty confident that it's unbiased. If you ask somebody else to flip a coin 200 times, and you can't actually see them, and you know they're lazy, and they come back and report exactly 100/100, that's a good indicator they didn't do the flips.
> It's suspicious when it lands on something that people might be biased towards.
Eh, this only makes sense if you're incorporating information about who set up the experiment in your statistical model. If you somehow knew that there's a 50% probability that you were given a fair coin and a 50% probability that you were given an unfair coin that lands on the opposite side of its previous flip 90% of the time, then yes, you could incorporate this sort of knowledge into your analysis of your single trial of 200 flips.
You can certainly do the frequentist analysis without any regard to the distribution of coins from which your coin was sampled. I’m not well studied on this stuff, but I believe the typical frequentist calculation would give the same results as the typical Bayesian analysis with a uniform prior distribution on “probability of each flip being heads.”
Even if you shoot only once, you still have a higher chance of hitting something slightly off the middle than the perfect 100/100. And this because that's one point-precise result (100/100) vs. a cumulated range of individually less-probable results, but more probable when taken as a whole.
> you still have a higher chance of hitting something slightly off the middle than the perfect 100/100
That's because "something slightly off the middle" is a large group of possible results. Of course you can assemble a group of possible results that has a higher likelihood than a single result (even the most likely single result!). But you could make the same argument for any single result, including one of the results in your "slightly off the middle" group. Did you get 97 heads? Well you'd have a higher likelihood of getting between 98 and 103 heads. In fact, for any result you get, it would have been more likely to get some other result! :D
> But you could make the same argument for any single result
Isn't that the point? The odds of getting the "most likely result" are lower than the odds of getting not the most likely result. Therefore, getting exactly 100/100 heads and tails would be unlikely!
But as I said, getting any one specific result is less likely than getting another other possible result. And the disparity in likelihoods is greater for any one specific result other than the 50% split.
I think the disagreement is about what that unlikeliness implies. "Aha! You got any result? Clearly you're lying!"... I'm not sure how far that gets you.
There's probably a dorm-quality insight there about the supreme unlikeliness of being, though: out of all the possible universes, this one, etc...
Try thinking of it this way: you're in highschool and your stats teacher gives you a homework assignment to flip a coin 200 times. You respect her and don't want to disappoint her, but at the same time the assignment is pointlessly tedious and you want to write down a fake result which will convince her you actually did it.
A slightly imperfect split is more likely to convince your teacher that you did the assignment. Intuitively this should be obvious.
> "Remember, if you flip a coin 200 times and it comes heads up exactly 100 times, the chances are the coin is actually unfair. You should expect to see something like 93 or 107 instead".
Inverting the statement makes it read something like this:
You are more likely to not get 100/100 than you are to get exactly 100/100
...which is exactly what I was saying. Nobody is arguing that there is a single value that might be more likely than 100/100. Rather, the argument is that a 100/100 result is suspiciously fair.
In statistics, various examples (e.g., coin flips) often stand in for other activities which might prove expensive or infeasible to make repeated tries of.
For "coin flips", read: human lives, financial investments, scientific observations, historical observations (how many distinct historical analogues are available to you), dating (see, e.g., "the secretary problem" or similar optimal stopping / search bounding problems).
With sufficiently low-numbered trial phenomena, statistics gets weird. A classic example would be the anthropic principle: how is it that the Universe is so perfectly suited for human beings, a life-form which can contemplate why the Universe it so perfectly suited for it? Well, if the Universe were not so suited ... we wouldn't be here to ponder that question. The US judge Richard Posner made a similar observation in his book "Catastrophe: Risk and Response" tackles the common objection to doomsday predictions that all have so far proved false. But then, of all the worlds in which a mass extinction event has wiped out all life prior to the emergence of a technologically-advanced species, there would be no (indegenous) witnesses to the fact. We are only here to ponder that question because utter annihilation did not occur. As Posner writes:
By definition, all but the last doomsday prediction is false. Yet it does not follow, as many seem to think, that all doomsday predictions must be false; what follow is only that all such predictions but one are false.
-Richard A. Posner, Catastrophe: Risk and Response, p. 13.
Well, yes. But the expected deviation from the mean is still ≈7.07. And the probability that the outcome will be either 93/107 or 107/93 is (slightly) higher than the outcome being exactly 100.
So while those are two results (93/107 and 107/93), they really only count as two separate outcomes if you pre-specify that the first number is heads.
If instead you consider symmetries, where there are 2 ways to get 93/107 and only one wall to get 100/100, then you have more likelihood for the 93/107 outcome because you have two ways to get it.
I don't think that makes 100 / 100 the most likely result if you flip a coin 200 times. It's not about 100 / 100 vs. another single possible result. It's about 100 / 100 vs. NOT 100 / 100, which includes all other possible results other than 100 / 100.
Depends what you count as a result, I guess. "There is exactly N flips of a single kind" is also a viable definition, just as "The exact sequence of flips was x_0, x_1, ... x_199" is.
Why not go one abstraction further and go expected deviation from deviation? Probably the word "expected" plays a mind trick? "Expected" doesn't mean the probability increases, the easiest way to understand it is just by looking at the probability distribution function chart for coin tosses - you'll immediately see that mean has the highest chance of happenning, so exactly 100 is the most likely outcome
I would guess that a single trial of 200 flips can be treated as one event, so getting 100/100 is but one outcome. It may be the most likely individual outcome, but the odds of getting that exact result feel less likely than all of the other possible outcomes. The 100/100 case should be seen the most over repeated trials, but only marginally over other nearby results.
Intuitively, this seems right to me, but sometime statistics do not follow intuition.
"fair coin" refers to both the probability of heads and tails being equal (which is still justified) as well as the trials being independent (unlikely with 100/200; more likely the "coin" is some imperfect PRNG in a loop)
You can use combinatorics to calculate the likelihood. If your PRNG is in a cycle of length N in its state space (assuming N>200), and half the state space corresponds to heads (vs tails), then the likelihood would be (N/2 choose 100)^2/(N choose 200) versus your baseline likelihood (for a truly random coin) of (200 choose 100)/2^200.
Graphing here https://www.wolframalpha.com/input?i=graph+%28%28N%2F2+choos... and it does look like it's only a slight improvement in likelihood, so I did overstate the claim. A more interesting case would be to look at some self-correcting physical process.
If a student was tasked with determining some physical constant with an experiment and they got it exactly right to 20th decimal place - I'll check their data twice or thrice. Just saying. You continue believing it was the most likely value ;)
The chance of exactly 100 heads from 200 fair coin flips is approximately 5-6%. Qualitatively, that's not particularly strong evidence for an unfair coin if you did only one trial.
You could also argue that 100 out of 200 on a fair coin is more likely than any other specific outcome, such as 93/200, so if the argument is that the coin is "too perfect", you then also have to consider the possibility that the coin is somehow biased to produce 93/200 much more often than anything else, vs. 100/200.
In a real-world scenario, if you saw a result significantly far from 100 (like 150 heads), you might suspect the coin is unfair. However, seeing exactly 100 heads gives no reason to suspect the coin is unfair; it's the result most consistent with a fair coin.
So many other good details that get to how impossibly multivariate biology research is, like the need to have several days at the exact temperature.
It's not uncommon for results in biology to have this kind of snag in reproducibility even now. Sometimes it's due to attributing variations to something like "steady hands at the bench", but other times it can even be a deliberate attempt to prevent rivals from duplicating a process before it can be patented and privatized.
Hare's theory predicts that there would need to be a cold snap at just the right time, and lo and behold there was. Probability isn't an issue if the only reason you are considering the probability is because the event already happened. Indeed the low probability of such an event transpiring goes a long way towards explaining why the discovery was not made earlier.
Root-Bernstein's theory makes no such testable predictions, and it solves the issue of an incomplete record on September 3rd with incomplete or inaccurate records elsewhere. It seems to me extremely plausible that fleming did not record the results of a botched, uncontrolled experiment but still recognized it as an indicator of something interesting that warranted follow-up. If I were in his position I would preserve the random dish for comparison to the more rigorous follow up experiment. I certainly don't put any stock into the argument that if the story had gone as Root-Bernstein describes it would have been too circuitous for scientific publishing, if anything it would be much more harmonious with standard scientific writing than the chance observation story.
- Fleming lived next door to an unsecure mycology lab.
- The temperature during the time period was low enough that if Fleming had left a contaminated culture unattended and non-incubated, he would have had a very high chance of getting the results he became famous for...
Well, given that the probability of discovering penicillin in those conditions is pretty high (say, if he forgot/neglected to incubate one out twenty batches, a 5% chance), and the prior probability of discovering penicillin any other way is extremely low (otherwise other scientists would have found it), bayesian calculus says the stroke of luck hypothesis is probably correct.
Preserving biological/medical specimens with formaldehyde so that they may be re-examined later is a common practice that still occurs today. So it isn't surprising here.
The alternative would be to freeze the specimens, but that is more cumbersome than stacking formaldehyde-fixed plates on a shelf, and specimens may alter upon repeated freezing and thawing.
One of my favorite "myths" about the discovery of stainless steel:
Metallurgist is trying out all kinds of steels looking for a particular attribute. He would dutifully record each recipe + test in a notebook but if a particular batch didn't have the attribute, he would throw it out a window into an outdoor scrap pile.
Several months go by and he's cleaning up the pile and notices that one of the blocks has no rust or corrosion. He knows that the pile is six months old but doesn't know which of the recipes this block was connected to.
So he repeats ALL of the block recipes from the last 6 months but labels each block so he can figure out which recipe led to the "stainless" steel.
I recall reading that the microwave oven was invented by a physicist after he walked by a radiation chamber and the chocolate bar in his pocket melted... makes me wonder if there was any historic license taken in that case as well.
The trick to actually melting chocolate in a microwave is to do it in pulses of only a few seconds, see if it stirs yet, and repeat. Longer runs will quickly destroy it.
Hmm, sounds like the spores must have come from the outside. Otherwise he'd be saying his colleague has contaminated the building with improperly stored fungal colonies and he himself let those spores contaminate his lab. So yeah, definitely from the outside.
"Presenter[John Cleese]: Penguins, yes, penguins. What relevance do penguins have to the furtherance of medical science? Well, strangely enough quite a lot, a major breakthrough, maybe. It was from such an unlikely beginning as an unwanted fungus accidentally growing on a sterile plate that Sir Alexander Fleming gave the world penicillin. James Watt watched an ordinary household kettle boiling and conceived the potentiality of steam power. Would Albert Einstein ever have hit upon the theory of relativity if he hadn't been clever? All these tremendous leaps forward have been taken in the dark. Would Rutherford ever have split the atom if he hadn't tried? Could Marconi have invented the radio if he hadn't by pure chance spent years working at the problem? Are these amazing breakthroughs ever achieved except by years and years of unremitting study? Of course not. What I said earlier about accidental discoveries must have been wrong. "
“Despite this close professional association, however, Hare claims to have played no part in the discovery or original research on penicillin nor to have discussed them with Fleming”
It’s nice to see that the bickering about who stole whose research does not affect all old discoveries.
Is it common in cases for someone with no involvement at all to claim involvement? Usually disputes I've heard of are when multiple people are involved, and they're arguing about who played a crucial vs minor role.
It's common for people who have close associations to have events that could be construed as involvement, and when someone does believe they are involved in something important, they tend to claim their involvement was important. It would be so easy to inflate a random conversation or a little common courtesy assistance as something more. It takes some genuine humility to take stock of all your interactions and conclude that you had nothing to do with one of the most important discoveries in history, and more still to admit that you thought nothing of it at the time.
The Flemming/Florey story is actually an interesting bit of science history, but it's about popular media and the romance of science v.s. how real science is done with patience, selfless dedication and a literature search. When Chain joined the Oxford lab, a literature search turned up 6 candidate "penicillins" which they went on to investigate. Fleming smoked a pipe, got to the office about 10, and was accessible (being in London). This version of the story admittedly comes from a Howard Florey biography I read 10 years ago..
For me the biggest plot hole of the popular penicillin discovery myth is clearly the open window story, including its multiple side plots with meteorological events.
And that that story exists and has PREVAILED despite the fact that everyone knows that the same building was housing a mycology laboratory only a flight of stairs away, and that it was the mycology colleague La Touche who identified the penicillin.
Tangentially related: doxycycline helps improve muscle and tendon tear recovery by enhancing the performance of matrix proteins that form "scaffolding", by inhibiting factors that break them down. "By inhibiting MMPs, doxycycline helps preserve and remodel the ECM, accelerating repair and improving biomechanical strength (e.g., tensile strength and reduced creep/strain)". Crazy. I'm a big fan off high-ROI off-label uses of well-tolerated, cheap, out-of-patent "WHO essential medecines list" pharmas. There's much to be found there.
Another tidbit: inderol (propranolol, beta-blocker) can aide PTSD recovery by reducing the emotional potency of traumatic memories when taken in a therapeutic replay.
That was a lot of words to get to the point that Fleming probably misremembered the sequence of events when he retold the story 15 years later. He even mentioned this possibility at the time. Interesting article but not much of a mystery.
And add how your method/insight moves the conversation forward, along with describing your method/insight.
The reason why scientific writing can be hard to read from the outside is the 100:1 compression. Decompression of that can require some knowledge of the field.
The goal of a paper is to write up the actual discovery, not to tell a story or explain irrelevant background. The steps he wrote there would confuse rather than elucidate the actual discovery.
In my opinion, we're possessed by a cultural epidemic of think pieces doing rich and nuanced science history, but wrongly framed in the form of correcting "myths" that, in their substance amount to quibblings over narrative emphasis. It's easy to get taken in by the framing because it truly is enlightening, and the argument goes down so smooth because its embedded in a rich, curious, and fascinating scientific history that otherwise embodies best practices I would happily celebrate.
But the key details about the story of penicillin are that a moldy plate showed bacteria-free clearing, Fleming saw it, isolated the mold, proved its germ-killing filtrate and published the finding, which is the heart of the story and which is not a myth.
I'm sure it's true enough that St Mary's windows were usually kept shut to keep pathogens in and contaminants out, that London's August 1928 cold snap would have slowed staph growth, that Fleming's first notes Or 8 weeks later than the actual event, and that a modern plate seeded with bacteria first will not produce the celebrated halo unless the mold is given a head start. The article makes much of the fact that today’s researchers cannot reproduce the famous halo if they add staph first, yet that difficulty rebuts a sequence Fleming never claimed to have used.
These points are significant, even fascinating, yet the article inflates them into a strobe-lit "MYTH" banner, turning normal human imprecision about times and temperatures into evidence of wholesale fiction, which abuses the ordinary friction of any retrospective account and punishes the story for the very human messiness that makes it instructive.
The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
It's a myth in the most literal way. Fleming published and promoted his results despite a lack of reproducibility. By the time he won the Nobel Prize, he had backformed or misremembered a folksy story about an open window. That's textbook mythmaking.
It can both be fine to have a glib story to tell schoolkids and important to recognize that the actual intellectual process is messier and more complex.
I have now actually read Flemming's 1929 manuscript that first described penicillin [0]. It is a careful and well documented scientific report describing the action of penicillin on various species of bacteria, how to produce it, and some of its chemical properties. It describes how penicillin can kill bacteria isolated from the throats of nurses, and shows that it has low toxicity in mice, and is possibly safe for use in humans: "Constant irrigation of large infected surfaces in man was not accompanied by any toxic symptoms, while irrigation of the human conjunctiva every hour for a day had no irritant effect."
It is far from having a "lack of reproducibility" and in fact allowed others to quickly and accurately replicate his discovery.
https://pmc.ncbi.nlm.nih.gov/articles/PMC2048009/
The path to his discovery may have been difficult to replicate, but the fact that the mold could kill other bacteria was not, and was immediately replicated.
It just wasn't seen as relevant because, at the time, few people imagined its internal use in humans and it was instead seen more as a tool for other microbiologists and the like. The jump to "And then I see the disinfectant, where it knocks it out in a minute. One minute. And is there a way we can do something like that, by injection inside or almost a cleaning?" took quite some time.
> The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
This is the same conclusion as the article. IMO, the importance of challenging the myth is that it has hisotrically taken precendence over your (and the article's) conclusion.
FTA
> Fleming’s 1929 penicillin paper may have been written as a linear process, but that’s almost certainly not how the discovery occurred. And by eliminating these complicated twists and turns, Fleming inadvertently obscured what may be one of the most important lessons in scientific history: how combining a meticulous research program with the openness to branch out into new directions led him to Nobel Prize-winning success. Neither rigid plans nor the winds of chance are enough on their own; discovery requires both.
> The window quibble, the incubator gap, and the replication protocol do not touch the central, uncontested fact that chance contamination plus observational curiosity gave medicine its first antibiotic.
Personally for me, while less important, I really appreciate the investigation into the narrative.
I agree that the science is more important and the results don't care about the story.
The balance is that we don't need to go around correcting everyone, but knowing more about the details of the story is worth my time in reading this piece. I think the article strikes the right tone.
To be anal about being anal, the article doesn't preclude Fleming's account. It argues that it's unlikely, but countless highly improbable things are happening every second. On this topic somehow Ancient Egyptian poultices (and in cultures onward - though they are the oldest recorded account) even used moldy bread to treat bacterial infections, somehow stumbling onto genuine antibacterial aspects for an absurdly counter-intuitive treatment that has a real effect. However it was initially discovered back then, let alone replicated and confirmed, must have been through an unimaginably improbable series of events. Yet it happened. That's rather the story of humanity.
Definitely worded for clicks, but remember that "myth" doesn't mean "false story", it just means "story".
Myth absolutely implies false (or at least exaggerated) story elements. If it wasn’t, we’d call it history.
And you can tell - ‘the history of penicillin’ implies a very different thing than ‘the myth of penicillin’ eh?
Oh I really enjoyed this one.
Got a quick insight about how penicillin works: interferes with cell-wall building which is a destroy and recreate process by preventing the recreate part.
Got a quick view into the scientific process and communication: Fleming focused on the insight - penicillium kills staphylococcus - and left out the circuitous detail. This is important so that the big win here is very clear.
And got an insight into human nature and memory: Fleming didn’t tell the accidental contamination story until much later. It could possibly be even an idea someone else might have come up with which then took root in his mind (ironic haha!)
The communication aspect reminds me of Mendel’s far too perfect ratios for his pea plants. That kind of “repeat till difference clear” statistics would be decried today but perhaps that was to communicate rather than to determine.
And finally, I really enjoy reading about human process innovation because I think it’s a big factor in how Humanity grows. The lab notebook has to be some kind of star performer here - Fleming’s notes allow us to look back like this.
When I experiment with things, I naturally lean to keeping notes on my test protocol, observations, and results. But not because of some personal genius. It’s just the standard way I was taught as a child in our science labs.
I won’t claim to the rigor of a microbiology lab but even just the process notes help a lot, which is useful since I’m just testing molecules on myself.
If you are not familiar with more of Mendel or plant biology, he got extremely lucky in picking a two chromosome species. The next plant he picked had more than two chromosome types so he spent the rest of his life hitting his head against the wall - obvious to us but him not having a theory and expertise with microscopes to explain his pea results hampered him greatly beyond his initial pea plant studies.
What do you mean by a two chromosome species? A quick google says pea plants have 14 chromosomes. I only looked because I had never heard of a species only having two chromosomes. Do you mean the traits he was selecting for only had two alleles?
The traits he picked- hope I get this detail pedantically correct: had only two alleles, each allele had an obvious phenotype controlled mostly by that gene, the two phenotypes were binary (no intermediate "half-wrinkled-half-smooth"). and all segregated independently (different chromosomes, or far enough that the linkage was extremely weak). I remember my high school teacher speculating he inspected many different phenotypes and then reported his results on the final ones he picked where the results were nice.
Unfortunately, most modern genotypes and phenotypes in humans don't follow these patterns, and over the years, genetics devloped an entire vocabulary and physical model to explain them, although at a fairly abstract level. None of it made any sense to me so I follow the biophysics/molecular biology approach which tends to consider many more underlying physical details
There's a related story, https://review.ucsc.edu/spring04/bio-debate.html and https://review.ucsc.edu/spring04/twoversions.html which shows how different fields think about the genotype/phenotype relationship. Grinding up steering wheels to figure out how they work...
Yeah that was my understanding. Most things have multiple genes controlling them. Like there is no single eye or hair color gene.
Diploid. Peas have 7 pairs of chromosomes. Many types of plants have multiple (polyploid) copies of each chromosome, and you will not discover the classic Mendelian dominant-recessive pattern by studying them. Mendel lucked out by studying a diploid species. https://duckduckgo.com/?q=polyploid&ia=web
Right, my bad for wrong terminology as biology nor botany is not my speciality either. I was thinking along the lines of XY for humans, then the ordinary two row, two column chart used to teach the basics of Mendel with pea plants and dominant/recessive non polygenic traits in introductory biology classes.
It's a shame that Fleming misremembered his process of discovery and created a myth of accidental discovery.
I like the Root-Bernstein narrative more. That in the monotonous execution of routine experiments for something unrelated an unusual observation 'forced' them to discover penicillins antibacterial properties.
Not an accidental discovery by good fortune in a serendipitous sense. An accidental discovery of a brute force exhaustive search. The narrative of we spent months meticulously examining hundreds of samples is less romantic, but is one that supports the importance of funding scientific inquiry.
We won't make progress by hoping people leave culture plates out on window sills. We make progress when we fund meticulous exhaustive efforts of discovery.
> Mendel’s far too perfect ratios for his pea plants.
"Remember, if you flip a coin 200 times and it comes heads up exactly 100 times, the chances are the coin is actually unfair. You should expect to see something like 93 or 107 instead".
Isn't 100 / 100 the most likely outcome for a fair coin? Sure it's unlikely that you hit exactly that result, but every single other result is even less likely.
What I'm trying to say: if you get 100 / 100, that's not a sign of an unfair coin, it's the strongest sign for a fair coin you can get.
> every single other result is even less likely.
But the summed probability of the “not too far away results” is much higher, i.e. P([93, 107]\{100}) > P([100]).
So if you only shoot 100/100 with your coin, that's definitely weird.
Okay, but it doesn't make sense to arbitrarily group together some results and compare the probability of getting any 1 result in that group to getting 1 particular result outside of that group.
You could just as easily say "you should be suspicious if you flip a coin 200 times and get exactly 93 heads, because it's far more likely to get between 99 and 187 heads."
It's suspicious when it lands on something that people might be biased towards.
For example, you take the top five cards, and you get a royal flush of diamonds in ascending order. In theory, this sequence is no more or less probable than any other sequence being taken from a randomly shuffled deck. But given that this sequence has special significance to people, there's a very good reason to think that this indicates that the deck is not randomly shuffled.
In theory terms, you can't just look at the probability of getting this result from a fair coin (or deck or whatever). You have to look at that probability, and the probability that the coin (deck etc.) is biased, and that a biased coin would produce the outcome you got.
If you flip a coin that feels and appears perfectly ordinary and you get exactly 100 heads and 100 tails, you should still be pretty confident that it's unbiased. If you ask somebody else to flip a coin 200 times, and you can't actually see them, and you know they're lazy, and they come back and report exactly 100/100, that's a good indicator they didn't do the flips.
> It's suspicious when it lands on something that people might be biased towards.
Eh, this only makes sense if you're incorporating information about who set up the experiment in your statistical model. If you somehow knew that there's a 50% probability that you were given a fair coin and a 50% probability that you were given an unfair coin that lands on the opposite side of its previous flip 90% of the time, then yes, you could incorporate this sort of knowledge into your analysis of your single trial of 200 flips.
If you don’t have any notion of how likely the coin is to be biased or how it might be biased then you just can’t do the analysis at all.
You can certainly do the frequentist analysis without any regard to the distribution of coins from which your coin was sampled. I’m not well studied on this stuff, but I believe the typical frequentist calculation would give the same results as the typical Bayesian analysis with a uniform prior distribution on “probability of each flip being heads.”
> So if you only shoot 100/100 with your coin, that's definitely weird.
Not if you only try once.
Even if you shoot only once, you still have a higher chance of hitting something slightly off the middle than the perfect 100/100. And this because that's one point-precise result (100/100) vs. a cumulated range of individually less-probable results, but more probable when taken as a whole.
For a fair coin, hitting 100/100 is ~5%, vs. ~30% falling in [97; 103] \ {100}. You can simulate here: https://www.omnicalculator.com/statistics/coin-flip-probabil...
> you still have a higher chance of hitting something slightly off the middle than the perfect 100/100
That's because "something slightly off the middle" is a large group of possible results. Of course you can assemble a group of possible results that has a higher likelihood than a single result (even the most likely single result!). But you could make the same argument for any single result, including one of the results in your "slightly off the middle" group. Did you get 97 heads? Well you'd have a higher likelihood of getting between 98 and 103 heads. In fact, for any result you get, it would have been more likely to get some other result! :D
> But you could make the same argument for any single result
Isn't that the point? The odds of getting the "most likely result" are lower than the odds of getting not the most likely result. Therefore, getting exactly 100/100 heads and tails would be unlikely!
But as I said, getting any one specific result is less likely than getting another other possible result. And the disparity in likelihoods is greater for any one specific result other than the 50% split.
I think the disagreement is about what that unlikeliness implies. "Aha! You got any result? Clearly you're lying!"... I'm not sure how far that gets you.
There's probably a dorm-quality insight there about the supreme unlikeliness of being, though: out of all the possible universes, this one, etc...
Try thinking of it this way: you're in highschool and your stats teacher gives you a homework assignment to flip a coin 200 times. You respect her and don't want to disappoint her, but at the same time the assignment is pointlessly tedious and you want to write down a fake result which will convince her you actually did it.
A slightly imperfect split is more likely to convince your teacher that you did the assignment. Intuitively this should be obvious.
Let's look at the original quote:
> "Remember, if you flip a coin 200 times and it comes heads up exactly 100 times, the chances are the coin is actually unfair. You should expect to see something like 93 or 107 instead".
Inverting the statement makes it read something like this:
You are more likely to not get 100/100 than you are to get exactly 100/100
...which is exactly what I was saying. Nobody is arguing that there is a single value that might be more likely than 100/100. Rather, the argument is that a 100/100 result is suspiciously fair.
Should that be 25% for 97..193 excluding 100?
“[97; 103] \ {100}” means the interval [97; 103] without the set {100}; so no, still ~30%.
I'm sorry, but try what once? 200 flips once?
In statistics, various examples (e.g., coin flips) often stand in for other activities which might prove expensive or infeasible to make repeated tries of.
For "coin flips", read: human lives, financial investments, scientific observations, historical observations (how many distinct historical analogues are available to you), dating (see, e.g., "the secretary problem" or similar optimal stopping / search bounding problems).
With sufficiently low-numbered trial phenomena, statistics gets weird. A classic example would be the anthropic principle: how is it that the Universe is so perfectly suited for human beings, a life-form which can contemplate why the Universe it so perfectly suited for it? Well, if the Universe were not so suited ... we wouldn't be here to ponder that question. The US judge Richard Posner made a similar observation in his book "Catastrophe: Risk and Response" tackles the common objection to doomsday predictions that all have so far proved false. But then, of all the worlds in which a mass extinction event has wiped out all life prior to the emergence of a technologically-advanced species, there would be no (indegenous) witnesses to the fact. We are only here to ponder that question because utter annihilation did not occur. As Posner writes:
By definition, all but the last doomsday prediction is false. Yet it does not follow, as many seem to think, that all doomsday predictions must be false; what follow is only that all such predictions but one are false.
-Richard A. Posner, Catastrophe: Risk and Response, p. 13.
<https://archive.org/details/catastropheriskr00posn/page/13/m...>
>But the summed probability of the “not too far away results” is much higher, i.e. P([93, 107]\{100}) > P([100]).
That's true of every result. If you're using this to conclude you have a weird coin then every coin is weird.
Low kolmogorov complexity equals suspicious
Well, yes. But the expected deviation from the mean is still ≈7.07. And the probability that the outcome will be either 93/107 or 107/93 is (slightly) higher than the outcome being exactly 100.
But those are 2 results. 100 / 100 is more likely than 93 / 107 (or any other specific result) is what I'm saying.
You can tell the difference between
93 heads, 107 tails and
93 tails, 107 heads
but not between
100 heads, 100 tails and
100 tails, 100 heads.
So while those are two results (93/107 and 107/93), they really only count as two separate outcomes if you pre-specify that the first number is heads.
If instead you consider symmetries, where there are 2 ways to get 93/107 and only one wall to get 100/100, then you have more likelihood for the 93/107 outcome because you have two ways to get it.
I don't think that makes 100 / 100 the most likely result if you flip a coin 200 times. It's not about 100 / 100 vs. another single possible result. It's about 100 / 100 vs. NOT 100 / 100, which includes all other possible results other than 100 / 100.
If you flip a coin 200 times, and repeat that a billion times, 100/100 will be more likely than any other ratio.
It will be a bell curve with 100/100 as the peak
Depends what you count as a result, I guess. "There is exactly N flips of a single kind" is also a viable definition, just as "The exact sequence of flips was x_0, x_1, ... x_199" is.
Why not go one abstraction further and go expected deviation from deviation? Probably the word "expected" plays a mind trick? "Expected" doesn't mean the probability increases, the easiest way to understand it is just by looking at the probability distribution function chart for coin tosses - you'll immediately see that mean has the highest chance of happenning, so exactly 100 is the most likely outcome
I would guess that a single trial of 200 flips can be treated as one event, so getting 100/100 is but one outcome. It may be the most likely individual outcome, but the odds of getting that exact result feel less likely than all of the other possible outcomes. The 100/100 case should be seen the most over repeated trials, but only marginally over other nearby results.
Intuitively, this seems right to me, but sometime statistics do not follow intuition.
"fair coin" refers to both the probability of heads and tails being equal (which is still justified) as well as the trials being independent (unlikely with 100/200; more likely the "coin" is some imperfect PRNG in a loop)
> more likely the "coin" is some imperfect PRNG in a loop
"More likely"? How can you even estimate the likelihood of the coin being "an imperfect PRNG" based on a single trial of 200 flips?
You can use combinatorics to calculate the likelihood. If your PRNG is in a cycle of length N in its state space (assuming N>200), and half the state space corresponds to heads (vs tails), then the likelihood would be (N/2 choose 100)^2/(N choose 200) versus your baseline likelihood (for a truly random coin) of (200 choose 100)/2^200.
Graphing here https://www.wolframalpha.com/input?i=graph+%28%28N%2F2+choos... and it does look like it's only a slight improvement in likelihood, so I did overstate the claim. A more interesting case would be to look at some self-correcting physical process.
Bayesian vs frequentist in a nutshell :)
If a student was tasked with determining some physical constant with an experiment and they got it exactly right to 20th decimal place - I'll check their data twice or thrice. Just saying. You continue believing it was the most likely value ;)
A truly fair coin would never land heads or tails. It would land standing on it's edge, every single time.
The chance of exactly 100 heads from 200 fair coin flips is approximately 5-6%. Qualitatively, that's not particularly strong evidence for an unfair coin if you did only one trial.
You could also argue that 100 out of 200 on a fair coin is more likely than any other specific outcome, such as 93/200, so if the argument is that the coin is "too perfect", you then also have to consider the possibility that the coin is somehow biased to produce 93/200 much more often than anything else, vs. 100/200.
There is also no way to weight a coin that makes it more likely to get 100/200 than a fair coin.
In a real-world scenario, if you saw a result significantly far from 100 (like 150 heads), you might suspect the coin is unfair. However, seeing exactly 100 heads gives no reason to suspect the coin is unfair; it's the result most consistent with a fair coin.
So many other good details that get to how impossibly multivariate biology research is, like the need to have several days at the exact temperature.
It's not uncommon for results in biology to have this kind of snag in reproducibility even now. Sometimes it's due to attributing variations to something like "steady hands at the bench", but other times it can even be a deliberate attempt to prevent rivals from duplicating a process before it can be patented and privatized.
cf The Harvard Law: "Under controlled conditions of light, temperature, humidity, and nutrition, the organism will do as it damn well pleases."
Hare's theory predicts that there would need to be a cold snap at just the right time, and lo and behold there was. Probability isn't an issue if the only reason you are considering the probability is because the event already happened. Indeed the low probability of such an event transpiring goes a long way towards explaining why the discovery was not made earlier.
Root-Bernstein's theory makes no such testable predictions, and it solves the issue of an incomplete record on September 3rd with incomplete or inaccurate records elsewhere. It seems to me extremely plausible that fleming did not record the results of a botched, uncontrolled experiment but still recognized it as an indicator of something interesting that warranted follow-up. If I were in his position I would preserve the random dish for comparison to the more rigorous follow up experiment. I certainly don't put any stock into the argument that if the story had gone as Root-Bernstein describes it would have been too circuitous for scientific publishing, if anything it would be much more harmonious with standard scientific writing than the chance observation story.
Yeah, I don't see the huge improbability here.
Given that we know that:
- Fleming lived next door to an unsecure mycology lab.
- The temperature during the time period was low enough that if Fleming had left a contaminated culture unattended and non-incubated, he would have had a very high chance of getting the results he became famous for...
Well, given that the probability of discovering penicillin in those conditions is pretty high (say, if he forgot/neglected to incubate one out twenty batches, a 5% chance), and the prior probability of discovering penicillin any other way is extremely low (otherwise other scientists would have found it), bayesian calculus says the stroke of luck hypothesis is probably correct.
> I would preserve the random dish for comparison to the more rigorous follow up experiment
This would also explain why the dish was treated with formaldehyde for preservation, and why the dish still exists today.
Preserving biological/medical specimens with formaldehyde so that they may be re-examined later is a common practice that still occurs today. So it isn't surprising here.
The alternative would be to freeze the specimens, but that is more cumbersome than stacking formaldehyde-fixed plates on a shelf, and specimens may alter upon repeated freezing and thawing.
One of my favorite "myths" about the discovery of stainless steel:
Metallurgist is trying out all kinds of steels looking for a particular attribute. He would dutifully record each recipe + test in a notebook but if a particular batch didn't have the attribute, he would throw it out a window into an outdoor scrap pile.
Several months go by and he's cleaning up the pile and notices that one of the blocks has no rust or corrosion. He knows that the pile is six months old but doesn't know which of the recipes this block was connected to.
So he repeats ALL of the block recipes from the last 6 months but labels each block so he can figure out which recipe led to the "stainless" steel.
(Probably not the real story but always loved this telling of it. Actual Wikipedia history is here: https://en.wikipedia.org/wiki/Stainless_steel#History)
I recall reading that the microwave oven was invented by a physicist after he walked by a radiation chamber and the chocolate bar in his pocket melted... makes me wonder if there was any historic license taken in that case as well.
That one I could totally believe. Radar equipment + chocolate bar = very likely to occur.
The main problem with that one is chocolate bar + body heat also leads to melting.
Please compare a chocolate bar after 30s in your pocket and 30s in your microwave.
Are you in the microwave for 30s too?
The trick to actually melting chocolate in a microwave is to do it in pulses of only a few seconds, see if it stirs yet, and repeat. Longer runs will quickly destroy it.
Hmm, sounds like the spores must have come from the outside. Otherwise he'd be saying his colleague has contaminated the building with improperly stored fungal colonies and he himself let those spores contaminate his lab. So yeah, definitely from the outside.
Best take on serendipty versus effort in the history of science is, naturally enough, from Monty Python:
https://www.youtube.com/shorts/1Hu0f_ti9EQ
Text from http://montypython.50webs.com/scripts/Series_3/99.htm
"Presenter[John Cleese]: Penguins, yes, penguins. What relevance do penguins have to the furtherance of medical science? Well, strangely enough quite a lot, a major breakthrough, maybe. It was from such an unlikely beginning as an unwanted fungus accidentally growing on a sterile plate that Sir Alexander Fleming gave the world penicillin. James Watt watched an ordinary household kettle boiling and conceived the potentiality of steam power. Would Albert Einstein ever have hit upon the theory of relativity if he hadn't been clever? All these tremendous leaps forward have been taken in the dark. Would Rutherford ever have split the atom if he hadn't tried? Could Marconi have invented the radio if he hadn't by pure chance spent years working at the problem? Are these amazing breakthroughs ever achieved except by years and years of unremitting study? Of course not. What I said earlier about accidental discoveries must have been wrong. "
This stood out to me:
“Despite this close professional association, however, Hare claims to have played no part in the discovery or original research on penicillin nor to have discussed them with Fleming”
It’s nice to see that the bickering about who stole whose research does not affect all old discoveries.
Is it common in cases for someone with no involvement at all to claim involvement? Usually disputes I've heard of are when multiple people are involved, and they're arguing about who played a crucial vs minor role.
It's common for people who have close associations to have events that could be construed as involvement, and when someone does believe they are involved in something important, they tend to claim their involvement was important. It would be so easy to inflate a random conversation or a little common courtesy assistance as something more. It takes some genuine humility to take stock of all your interactions and conclude that you had nothing to do with one of the most important discoveries in history, and more still to admit that you thought nothing of it at the time.
The Flemming/Florey story is actually an interesting bit of science history, but it's about popular media and the romance of science v.s. how real science is done with patience, selfless dedication and a literature search. When Chain joined the Oxford lab, a literature search turned up 6 candidate "penicillins" which they went on to investigate. Fleming smoked a pipe, got to the office about 10, and was accessible (being in London). This version of the story admittedly comes from a Howard Florey biography I read 10 years ago..
If it's the Leonard Bickel one, it's excellent.
Fleming is a folk hero. Florey and Chain (and Du Pont, for industrialisation) did the real work.
imo, this paragraph covers the essence of a good chunk of the article: https://en.wikipedia.org/wiki/Discovery_of_penicillin#Replic...
For me the biggest plot hole of the popular penicillin discovery myth is clearly the open window story, including its multiple side plots with meteorological events.
And that that story exists and has PREVAILED despite the fact that everyone knows that the same building was housing a mycology laboratory only a flight of stairs away, and that it was the mycology colleague La Touche who identified the penicillin.
The Hare theory is a better story, regardless of whether it is true. I am surprised that it hasn't seeped into pop science lore.
Tangentially related: doxycycline helps improve muscle and tendon tear recovery by enhancing the performance of matrix proteins that form "scaffolding", by inhibiting factors that break them down. "By inhibiting MMPs, doxycycline helps preserve and remodel the ECM, accelerating repair and improving biomechanical strength (e.g., tensile strength and reduced creep/strain)". Crazy. I'm a big fan off high-ROI off-label uses of well-tolerated, cheap, out-of-patent "WHO essential medecines list" pharmas. There's much to be found there.
Another tidbit: inderol (propranolol, beta-blocker) can aide PTSD recovery by reducing the emotional potency of traumatic memories when taken in a therapeutic replay.
That was a lot of words to get to the point that Fleming probably misremembered the sequence of events when he retold the story 15 years later. He even mentioned this possibility at the time. Interesting article but not much of a mystery.
The cartoon 60% of the way down the article definitely feels like an "artists barely concealed fetish" thing.
(Off-topic:) Scrollbars, and their non-existence.
It's not the best implementation, but it's there (you just have to scroll a bit to get it to show).
I wish this story were not framed as a “myth”. Far fewer people would read it if they knew the (presumed) truth - a small deviation from the story.
Can someone with more experience in scientific writing comment on
> It’s too circuitous and indirect for a scientific report
The preceding paragraph does not seem unreasonable to me--maybe a bit too glib, but nothing that couldn't be touched up.
in scientific writing, you need to take
100 articles at 6 pages each
and condense them to 1 article of 6 pages
And add how your method/insight moves the conversation forward, along with describing your method/insight.
The reason why scientific writing can be hard to read from the outside is the 100:1 compression. Decompression of that can require some knowledge of the field.
Also, some people are just bad at writing.
The goal of a paper is to write up the actual discovery, not to tell a story or explain irrelevant background. The steps he wrote there would confuse rather than elucidate the actual discovery.
Not a mention of Clorito Picado?
https://en.wikipedia.org/wiki/Clodomiro_Picado_Twight
The contest is cold, but it deserves at least a fleeting hint
Writing like this does not belong on Substack.