Laboratory nightmare!

I remember my wild man days, when I might have as much as 5 or 6 cups of coffee during the day — to be honest, I was never particularly wild, but I would get the caffeine shakes and feel a bit edgy. I’m now down to one cup in the morning, and if I’m feeling crazy, two. But never have I consumed the equivalent of 300 cups of coffee in one sitting, like these unfortunate students.

The students had volunteered to take part in a test in March 2015 aimed at measuring the effect of caffeine on exercise.

They were given 30g of caffeine instead of 0.3g, Mr Farrer said.

Death had previously been reported after consumption of just 18g, he told the court.

The university had switched from using caffeine tablets to powder, he said.

“The staff were not experienced or competent enough and they had never done it on their own before,” he said.

“The university took no steps to make sure the staff knew how to do it.”

The calculation had been done on a mobile phone, with the decimal point in the wrong place, and there was no risk assessment.

Caffeine is a potentially dangerous drug, and anyone working with it needs to be aware of that fact. I’ve had students experiment with it in our cell biology lab, and I always preface providing the purified drug to students with the warning that it is almost certainly the most dangerous chemical in the lab at that time, that they shouldn’t let familiarity with it as an ingredient in coffee and soft drinks let them be casual with playing with it.

I probably looked like a gaffed fish while I was reading that. It was appallingly sloppy practice.

  • They were dosing students with this drug, not mice or some other animal model.
  • They didn’t have competently trained staff monitoring every step.
  • They changed the reagent from over-the-counter pills to purified powder, which ought to have had everyone triple-checking the concentration.
  • They didn’t have their protocol vetted by an experienced pharmacologist.
  • They relied on a calculator.GIGO. I’ve been nagging my students to do more estimation, because I’ve noticed that calculator-dependent students easily make errors that are many orders of magnitude off, and they are completely unaware.

In our introductory labs, when we have students calculate concentrations, we have a little check box in the write-up — they have to get a TA or instructor to sign off on the calculations, and that’s for safe procedures, like adding an indicator dye to a tube of yeast. Every year when I’m grading lab reports, I keep an eye open for egregious errors in concentrations. One year the record was set by someone being off by 31 orders of magnitude.

I know students are easily capable of bone-headed errors in arithmetic that they aren’t experienced enough to notice, and to do it in an experiment in which students are the subjects…no. Just no. Now I’m going to have nightmares.

At least in this case the two poisoned students survived, and the university was fined £400,000. They all got off easy.


  1. Artor says

    I am reminded of a nuke plant disaster in Japan in 1999. From Wikidepia:


    blockquote>The accident occurred as three workers, Hisashi Ouchi, Masato Shinohara, and Yutaka Yokokawa,[9] were preparing a small batch of fuel for the Jōyō experimental fast breeder reactor, using uranium enriched to 18.8% with the fissile radionuclide (radioisotope) U‑235 (with the remainder being the fissionable-only U‑238). It was JCO’s first batch of fuel for that reactor in three years, and no proper qualification and training requirements appear to have been established to prepare those workers for the job. At around 10:35, a precipitation tank reached critical mass when its fill level, containing about 16 kilograms (35 pounds) of uranium, reached about 40 liters (11 U.S. gallons).[8]

    Criticality was reached upon the technicians adding a seventh bucket of an aqueous uranyl nitrate solution to the tank. The nuclear fission chain reaction became self-sustaining and began emitting intense gamma and neutron radiation. At the time of the event, Ouchi had his body draped over the tank while Shinohara stood on a platform to pour the solution into it; Yokokawa was sitting at a desk four meters away. All three technicians observed a blue flash (possibly Cherenkov radiation) and gamma-radiation alarms sounded.[7] [10]

    Technicians Ouchi and Shinohara immediately experienced pain, nausea, difficulty breathing, and other symptoms. Ouchi then began to vomit in the decontamination room a few minutes later and lost consciousness shortly after.[11] Fission products such as yttrium‑94 and barium‑140 began contaminating the building.

    Being a wet process with an intended liquid result, the water sustained the chain reaction by serving as a neutron moderator, whereby neutrons emitted from fissioned nuclei are slowed so they are more readily absorbed by neighboring nuclei, inducing them to fission in turn. The criticality continued intermittently for about 20 hours. As the solution boiled vigorously, steam bubbles attenuated the liquid water’s action as a neutron moderator (see Void coefficient) and the solution lost criticality. However, the reaction resumed as the solution cooled and the voids disappeared.

    The following morning, workers permanently stopped the reaction by draining water from a cooling jacket surrounding the precipitation tank. The water was serving as a neutron reflector. A boric acid solution (boron selected for its neutron absorption properties) was then added to the tank to ensure that the contents remained subcritical.[8]

  2. says

    If caffeine is the most dangerous thing in the lab you’re not doing it right. As an apprentice I was once given 100g of KCN and a lab where I worked unsupervised most of the time.

  3. Artor says

    Deadly and dangerous are different things. Who is going to be careless with potassium cyanide?

  4. cates says

    “Who is going to be careless with potassium cyanide?”
    Somebody; somewhere:
    ‘Hold my beer’

  5. nomaduk says

    When I was in high school Chemistry, our teacher, Mr Callahan, prohibited the use of calculators, requiring instead that all students obtain and learn to use a slide rule, because (a) using a slide rule requires that you have some idea of the magnitude of the result beforehand and (b) they were good to 3 significant figures, which was generally more than enough for what we were doing. I haven’t seen anything in the years since to convince me that his approach was in any way wrong.

  6. nomaduk says

    And, in fact, in my job as an IT support technician for the local school district I see evidence every single day that he was absolutely right.

  7. hemidactylus says

    I am typically hopped up on coffee. Once in a while I will mainline energy drinks. In the distant past I took Vivarin and met my limits. I also messed with ephedrine pills way back when. Not good. Bad side effects including crawling hair shafts. Sometimes, due to colds, I will take Claritin D, the pharmaceutical equivalent of a speedball. Body gets confused by antihistamine drowsiness and pseudoephedrine hype. Weird loopy feeling.

    Anyway I sometimes drink coffee excessively but only occasionally amp up on a rush of Cuban coffee, the beast of espressos.

    But I doubt my consumption ever comes close to what happened to these students. Yipes!

    By the same token I prefer beer to liquor. Volumetrics. I feel bloated and the little satiation signal goes off in my brain that doesn’t come with doing multiple shots in quick succession.

    Now some of them hearty stouts mix the coffee and beer experience. Not same as those crazy high octane beers that went full energy drink level by adding caffeine to bad effect. Stout is more about the taste than physiological effect.

  8. Kevin Karplus says

    I’ve never had a student make a mistake bigger than 12 orders of magnitude (mixing up M and µ modifiers), but I’ve been teaching electronics with low voltages that are not particularly dangerous, even in the hands of idiots.

    The dosage error on caffeine with no one competent checking the work is terrifying.

  9. says

    I’ve been toying with the idea of banning calculators in cell biology, and instead make them build slide rules. You can download paper slide rules now, just print ’em out, a little assembly, and go.

    I just don’t know if I want to deal with the emotional breakdowns in class.

  10. garnetstar says

    Oh goodness, how many times have I told chem students that, if they don’t learn unit conversions, they’ll one day administer grams instead of milligrams, or milligrams instead of micrograms, and kill someone? I didn’t know that it’d be this literal. And, sorry to say, we are mandated to let them use calculators.

    Oddly, in chem labs, it is the daily and common chemicals and supplies that usually are behind the worst accidents, even the deaths. As @3 said, if it says DANGER, etc., on the bottle, at least you notice. But if it’s caffeine, you don’t try or even care.

    In my lab we work with very dangerous chemicals, pyrophorics, toxins, and all that, but the one I am afraid of is oxygen gas. I have to warn all the students and many professors, who all think that oxygen is just beneficial, how many people it has killed and maimed.

    Gasoline is the only other chemical I’m afraid of: my next car will be electric if only so that I don’t ever have to breathe gasoline or be anywhere near it, ever again.

  11. Hoosier X says

    Artor @ #1

    Where was the safety supervisor?


    I was getting donuts, sir!

    Aww, don’t be mad at Homer! Those pink donuts with the speinkles are hard to resist!

  12. whheydt says

    The worst failure to use a calculator example I ever–personally–encountered was rather cool, and not excessively (though obviously–to me at least–wrong).

    There was a guy who would stand in a hallway at SF cons with a 12″ globe of the Earth at his feet. He would hand a globe of the Moon to the same scale to people and challenge them to put it down the correct, scale distance, away. He was startled when I paced off 30 feet. I gave him both the real numbers, Earth 8K miles in diameter, Moon 240K miles away, and the shorthand way to do the scaling: the Moon is 30 diameters from the Earth.

    He had though the correct separation was 10 feet, though he conceded that I was correct. He claimed that the batteries were running down in his calculator when he did the original determination. I think he missed a factor of Pi.

    As for slide rules in high school… I’ve felt that way for decades for just the reasons given. You have to know the order of magnitude before you start and a rough idea of what a reasonable answer is. I went though high school using a slide rule (a circular one much of the time) and then through college as well. My main supplement was a set of 5 place log tables (CRC “Rubber Handbook” math tables…the older small format version). Of course, at my age, that was necessity, not virtue. When I was in college, a 4 function hand held calculator cost $400, which was way out of the price range for all but a very few students. By comparison, room and board in the dorm I was in at Berkeley was $920 per year.

  13. chris61 says

    @6 @7

    I grew up before calculators and I couldn’t agree more. Slide rules are definitely a useful learning tool that I wish more teachers and undergrad professors used. It’s kind of frightening to contemplate the number of graduate students working in laboratories who can’t estimate what a correct calculation should look like.

  14. blf says

    Another olde here who was taught how to use a slide rule in high school chemistry (albeit I already broadly knew how due to my dad). Whilst I don’t use them anymore the skills in estimating order-of-magnitude (and keeping track of units!) have proved invaluable in the decades since. Also quick numerical estimation, along the lines of @15.

  15. Pierce R. Butler says

    Caffeine has very potent effects in even small quantities for those not accustomed to its use. I gave it up decades ago after realizing I had become dependent on it, and not liking that on general principles.

    A while later, I set off to climb a local mountain, but failed to find the spring friends had directed me towards, and came down seriously dehydrated. I met some friends whose only available beverage was some Coke, which I desperately chugged (it felt like my parched throat lining absorbed all of the first swallow).

    The friends were on their way to cut up a couple of dead junipers for firewood (conscientious desert dwellers do NOT cut live trees except in direst emergencies). They went at one with axes and saws, and I did the other with my bare hands. I finished first. Do not underestimate caffeine!

  16. square101 says

    I feel like the main issue in this situation isn’t the calculation issues, but an issue with the fact that the people administering the experiment were ignorant, careless and uneducated about what they were doing. If they had any idea about caffeine doses they could have made 100s of mistakes with both calculators and slide rules and still looked at “30 g” of caffeine and I known that that was a crazy unsafe dosage and fixed the error or at least talked to someone who knew what they were supposed to be giving people. I realize that people who grew up with more demanding tasks being required to learn math often lament the use of calculators but I guarantee you that forcing students who have never users a slide rule before will not make them magically more carefull and exacting in their math, in the same way that forcing your students to write their papers with a type writer will not magically produce better manuscripts.

  17. says

    If I hope to teach my students one thing in class it’s how to do estimates and get a feeling for numbers, but so far with little success. One day they’ll be cheated out of their money because they simply can’t estimate what 2487 X 49 is.

  18. cartomancer says

    Hmm, I’ve heard the term “slide rule” used before, in the context of antiquated mathematical equipment, but I’m not sure I could honestly say I’ve ever seen one, much less tell you what it’s supposed to do. Is it like a cross between a protractor and a set of Napier’s rods, or more like an Astrolabe?

  19. Mobius says

    Teaching calculus, there have been numerous times I’ve had students insist the answer had to be right because that is what the calculator said.

  20. Sili says

    Orders of magnitude isn’t the only way to fuck up. I vaguely recall some poor nurses mixing up drips of potassium and sodium chloride (kalium and natrium) because the bags weren’t in any obvious way distinguishable.

  21. blf says

    Speaking of ways to feck up, one that has annoyed me for decades is the output from (generally technical system) programs being opaque as to the radix (hexadecimal, decimal, octal, binary, or…) or units. Is “size = 123” 123, 291, 83 bytes, kibibytes, or… or what… memory, data consumed, data generated,…? It’s infuriating. And can — and has — led to mistakes and errors.

    Having said that, what may be my favourite all-time computing error had to do with summer time (daylight saving time), and the switch to/from normal time. There was an automated steel(?) works in Germany which cooked its whatever for so-and-so many hours. The control program measured that by monitoring the elapsed local time; e.g., a four-hour cooking started at 01:00 ends at 05:00. Yes, but… not if, just if, there happens to be a change from Summer to normal time, then an extra hour is added. Which there was. So the stuff cooked for an extra hour. It burst through the vat and flooded the factory.

    Fortunately, being completely automated, there was no-one on the factory floor, so no-one died (or was injured?). However, with tonnes of now-solidified molten steel(?) covering everything, the factory was a write-off.

    (From a system analytics view, one has to wonder why there apparently weren’t temperature sensors and other safeguards. It’s possible there were and the story is more complicated than the above (admittedly vague) synopsis… and I do wonder what one would do with a vat full of solidified molten steel, which seems like a possible outcome of a “failsafe” cutoff. (For “vat” you should probably read “blast furnace”?))

  22. blf says

    me@27 (an edit seems to have gone missing?): Is “size = 123” 123, 291, 83 bytes, kibibytes, or… → Does “size = 123” mean 123, or 291, or 83, and is it bytes, kibibytes, or…

  23. blf says

    The point is not insomuch about using a slide rule, but gaining a feeling for what a plausible, and perhaps more importantly, impossible, result looks like. Consider the example at @15: A one-foot diameter Earth would have (to scale) a Moon about 30 feet away. No slide rule, or calculator, needed, not even knowledge of the relevant mathematics, just a pair of facts and some quick mental approximate calculation.

    The people who have used slide rules seem to tend to have this feel. Of course it takes time to learn to use a slide rule, but you’re learning more than how to manipulate a slider and scale. It’s that “more” which is a critical thing here! Slide rules more-or-less force that on you — you just don’t get plausible answers (except by guessing) until you have some functional grasp of the other details.

    A (possible?) alternative to banning calculators is to require the calculations to be written out, including units and, when applicable, rationale / formulae. That doesn’t mean doing the calculation by hand — or by slide rule — but does give the student an extra opportunity to spot an error. (And flag that they have made an error, even if they can’t find it.)

  24. says

    The clean joke Gershon Legman couldn’t resist including in Rationale of the Dirty Joke (volume 1) is about slide rules.

    The professor is lecturing. “Today, class, we will learn the operation of the slide rule, by which any two numbers can be multiplied together with admirable precision. Please, somebody furnish two numbers for the demonstration.”

    The inevitable voice from the back of the room shouts out “TWO TIMES TWO!” and, without missing a beat, the professor continues. “Simply move the 1 of your C scale so that it is aligned over the 2 of the D scale. Now move the hairline to the 2 of the C scale, and read the answer from the D scale. As you will see, the answer is three point nine, nine… eight. But we’ll call it four.”

  25. blf says

    Kip TW@31, That reminds me of a coworker with my father, who dad said was the only person he ever knew who would read off five or six places from a slide rule — and insist they were correct.

    A vaguely related related story from one place I used to work, a “study” done by the documentation experts asserted that using such-and-so networked printing system would mean printing at “approximately 1234.567 bytes per second” (value paraphrased) — um, approximately, to an implied thousandth of a byte? (Also a bit of useless number in any case, as either glyphs or pages per unit time would be more meaningful.)

  26. says

    My digital slide rule would have produced the necessary accuracy! The slider would have a digital LED readout. (No, it never existed, or at least mine was just a drawing. Though I do have an eight-foot slide rule here that would produce some accuracy if used correctly.)

  27. magistramarla says

    I’m a humanities person, not a math and science person, but I thought that learning to use the slide rule in high school science class was fun. My chemistry teacher brought in his very expensive calculator one day and let us play with it a bit. He told us that he predicted that as the calculators became less expensive, they would replace slide rules. He was certainly correct.

  28. photon says

    Back in the last century, when I was an undergraduate, one of the local universities had a bit of an oops. They were conducting an experiment that involved injecting subjects (students) with very small amounts of a radioactive tracer in order to assess different first aid treatments for snakebite. Unfortunately, someone misplaced a decimal point, someone else didn’t check the calculation, and the students ended up getting a dose several orders of magnitude greater than intended.
    I believe the final settlement included the university (or their insurers) paying for all the victims’ medical expenses for the rest of their lives.

  29. khms says

    I remember using a slide rule in school. I also remember that it was over complicated – it had a lot of functions I never figured out what they were for. And it wasn’t much of a speed improvement on doing it by hand or in my head, especially since for estimation purposes (so I’d know where to put the decimal point) I’d have to do much of the separate calculation anyway. I don’t think I ever did anything with it after school (except using it to draw lines), and good riddance.

    Yes, people need to learn to estimate things. But I don’t think a slide rule is all that helpful with that, and it takes serious time really learning how to use one. It is a distraction.

    Perhaps presenting students with pre-calculated results and having a speed test (how many problems can you answer correctly in a given short time) where either you have several different magnitudes of results to chose from, or given a result have to decide if it’s correct. Hey, you could even make a piece of software to make these up! And tell them they need to pass a certain number (per time) to pass. Make it so using a calculator pretty much cannot be fast enough. Hmm, one could set up a web site for that …

  30. Knabb says

    The slide rule is a distraction – teaching numerical approximation done by head aiming for approximate answers is far more useful. Using Gilliel’s example of 2487 * 49, sure, you could bust out the slide rule and have to at least be roughly accurate. Or you could glance at that, see that it’s close to 2,500*50, and know that the end result should be in the vicinity of 125,000 and that if you get an answer of 1218.63 you’re clearly off by more than an OOM (exactly 2, in this case).

  31. jack16 says

    The late Isaac Asimov writing in F&SF made a huge error. A reader corrected him. Writing afterward he speculated that it might be a world record mistake. (It was something like forty orders of magnitude.)