As a doctoral researcher, one of the “perks” has always been babysitting the undergraduates. I don't use that phrasing ironically. As a postgraduate demonstrator, you're not so much leading a laboratory as watching eighteen and twenty year olds working independently to make sure they don't set anything on fire – or put it it out quickly if they do. This leads to a very condescending demonstrator mentality, embittered by the disappointing responsibility for, if anything goes wrong, staying there later, probably alone with the same cack-handed student who initiated the disaster.

And it isn't so much their lack of practical skills (although the number of perished rubber tubes strewn across heating plates might seem to suggest another theory), as lack of reading skills. For example, I had one student who decided to cut some corners when it came to rinsing out equipment by filtering into the same Buchner flask twice and then throwing away the mixed waste from steps 1 and 2 of the lab. Unfortunately, he had failed to read beyond the line in step 2 which read “collect the product” to realise that whilst the product from step 1 was the solid, the product from step 2 was the liquid and would be needed for step 3. ...Which he had just combined with earlier waste. So, after a failed battle with the separating funnel, he was obliged to start all over again, steps 1 and 2 as well as 3. And during the last peaceful hour, whilst I sat on the far side of the lab bench watching him like a hawk, backed by the gentle music of his happily chuntering filtration, I explained to him why, even if he had been collecting a solid, not clearing out your Buchner flask was a bad corner to cut. A bit of suck back, and your solid ends up embedded in the sludge from earlier in the lab, contaminating it, or – the gut-wrenching horror of every undergraduate – messing with your yield. And that, I explained, is how you learn the most important skills you can acquire for practical chemistry: how to cut the right corners.

Of all the weird and wonderful theories about heredity, telegony, “offspring at a distance”, has to be the most interesting and horrendously misused. This is the singular theory that previous males who have impregnated a female will inflict their characteristics on her offspring from a current partner. Fathered by Aristotle and rediscovered with the re-emergence of his works in the Middle Ages, it's actually more of a mishmash of a couple of theories – prepotency – the idea that some individuals impress their characteristics on their offspring more effectively than others (a natural skill perhaps, like possessing better balance) – and maternal impressionability – the idea that the events experienced by a pregnant female would impress upon the foetus. These beliefs naturally excuse the monitoring and controlling the movements and actions of women and maternal impressionability is even found in the Bible, where Laban of the Book of Genesis produced striped lambs by showing pregnant ewes striped hazel rods.

Telegony is, of course, intimately tied with the patriarchal concept that females have control over pregnancy and childbirth and keep it a secret from males to disempower and annoy them (e.g. by harvesting their sperm and “keeping” it until they choose to conceive). Even Henry VIII was concerned about this possibility, which may explain why he murdered one of his wives and changed the religion of an entire country by force in order to rid himself of another. Neither is it dissimilar to modern Christian doctrines who say that a woman cannot become pregnant if she is raped. Some fundamentalists cults even propagate the theory of telegony now in order to enforce chastity via scare tactics.

On the flip side for women, telegony meant that in the case of an affair, the legal rather than biological father was considered to have a higher claim over any illegitimate children his wife might mother, so the fact she was sleeping around was never that important anyway.

Most of the theory was explored through pre-Mendelian discovery genetic exploration of mating horses first with zebras and then with other horses to see whether the horse-horse offspring were stripey or resistant to African diseases (it varied). Even Darwin toyed with telegony, and was convinced by this kind of stripey foal evidence as late as 1868. This led to a lot of animal breeders becoming excessively worried about the “contamination” of their females, whose taste was seldom as pure-bred as that of her breeders.

In fact, it wasn't until statistician Karl Pearson pointed out that telegony also implied an increasing similarity between father and successive children belonging to the same mother that the discrepancies overwhelmed the coincidences and telegony actually, at least outside fundamental Christian circles, disappeared.

Spontaneous generation is the idea that life can just pop into existence. Today, this sounds silly. It sounds like some Hand of God idea that the almighty looks down one day, thinks, that field could look a little prettier with just one more sheep in it and poof, he's made a new sheep.

But this is not about making sheep.

No, actually spontaneous generation is about conservation of matter. Matter in the form of living things doesn't just appear, existing matter from other things unites to build a new living thing. This is like anti-entropy (and probably does need some sort of divine inspiration to input the driving energy). The universe has an internal sense of order where life, "sweet, screaming, pooping life", is a higher order state. It was Aristotle's baby. Anaximander, Hippolytus and Anaxagoras were also pretty keen on it, and thought life would emerge from slime, mud and earth, so long as there was sunlight on it. This is actually a pretty cool and counter logical deduction. Then Pasteur had to go and spoil it by showing that meat in a sealed container didn't develop maggots.

This put a stop to the theory because if spontaneous generation had been hand of god derived, it wouldn't make much sense to say god could not get into your sealed box, i.e. he just can't think inside the box. But if the matter that is supposedly conserved is locked into the box, and no life sprouts, and in that which is not boxed up, life does sprout (and you repeat it many times until your statistics are convincing), you eventually have to come to the conclusion that life comes from outside the meat, not through conservation of the matter in the meat, and then all you need to do is develop a good enough microscope to see the fly eggs.

This was supposed to work. It always works - except on stage, before a crowd of interested onlookers, or when you're trying to capture it on camera. The rocket does not go off, the bubbles do not catch fire, the Newtonian fluid does not set, the light bulb does not bounce. They just sit there looking innocently at you whilst you put your hand through a Bunsen burner or into a big vat of custard and smash your props in front of the people who are excitedly waiting to be shown SCIENCE.

Because science doesn't work, that's how it does work. Science is just not 100%. That's why people who don't believe in its theories feel validated. And you have to admire the confidence of the Humphrey Davys, willing to expose themselves to this ridicule. Experiments fail. Especially when everybody is watching.

Is there a scientific explanation for this phenomenon?

Well, no - because every experiment and every example of every experiment is different, uniquely identified along the space time axes. But there are some general explanations that may go part way toward explaining why stuff keeps blowing up in your face - and very much not literally.

1. Your technique. Upon a stage, you will do things differently: sometimes with a flourish, sometimes with hyperbolic vigour - mostly, with over-caution, whether because you're experiencing stage fright, or because you don't want the overflow to end up on your audience. You press things harder, weaker, more suddenly - your hand shakes as you connect parts... and stuff is just that bit more likely to go pear-shaped.

2. Your equipment. If you're not presenting where you practised, you might have different equipment: a table with a dodgy leg, a less pure chemical, or a container with a leak in it. Unlike in your native environment, there also won't be a spare for you to reach for or five minutes for you to take out taping that hole up.

3. Pre-preparation. Because it's unlikely you will be doing an experiment that doesn't require setting up, which you will have done earlier. Probably off-site before transporting your equipment to the stage upon which it is set. It may have got jiggled around a bit. It may just have been left sitting for too long and developed problems such as leaks and tilts.

4. Sod's law. Because probability is just like that, and sometimes you can do everything right, but the time you care most about something working, probability may mean that's the time it all hits the fan, for no good nor foreseeable reason.

5. Psychology. This doesn't actually affect whether your experiment will work or not, it just affects whether you think it has. Before it becomes very confusing (the experiment is a Schrödinger cat: both successful and unsuccessful simultaneously!) - let me simplify: your experiment fails in practise runs all the time, but you don't remember it. It doesn't fail during performance any more often: these are just the ones you do remember.

So really, it's all your fault.

If you've read or watched an Alice in Wonderland story, you'll have heard of the Mad Hatter and seen his unusual behaviour. I hear the phrase “mad as a hatter” all the time, usually applied to myself. But at least I'm not suffering from chronic mercury poisoning.

Back in the eighteenth century, it was an occupational hazard of felt hat-making. And felt hats were, of course, all the rage.

Orange mercury nitrate was considered a necessary ingredient: it got smeared over the surface of the furs and shaved off once it had dried, allowing furs to be merged and the hairs to stand to attention. This was called carrotting. It worked pretty well. Really, it was a shame that mercury nitrate was a neurotoxin that severaly poisoned the hatters.

Sometimes this was through direct ingestion: whilst they were painting on the nitrate, the hatters would actually lick the tips of their brushes to sharpen the tips – much like painters who then got lead poisoning. Even if they didn't do this, however, the next stage of the process would allow them to breathe it in: as they shaved off the hardened mercury nitrate to produce the finished product, it would incidentally vaporise into a thin dust.

Mercury poisoning is not nice. Hatters suffered from all kinds of symptoms, ranging from confusion and emotional distress to reddening, shaking and muscular weaknesses. Eventually, it would probably kill them.

It wasn't until 1898, in France, that somebody realised it wasn't a good idea to allow people to do a job that slowly poisoned them, and passed a law to prevent it. The fad spread until mid twentieth century, when even the Americans stopped using mercury for hats.

It took a good deal longer for chemists in labs to stop picking it up in their hands and playing with it, of course.

If you say “DIY Chemistry”, you immediately conjure up images of children sticking mints into two litre bottles of fizzy pop, setting fire to various plastics in a mini bonfire in a pit they've dug into the pebbling in the back yard, or boiling fragrant petals in an attempt to distil their extracts. It's not that I didn't do all of these things as a child, it's that that isn't surprising. Kids. No matter how much everybody moans about the cotton wool generation, kids still have a natural tendency to put things in their mouths and set them on fire. Just not at the same time.

When I say “DIY Chemistry”, I'm talking about adults. Because despite the safe and broadly available tools for chemistry and other science experiments that exist in the world today, there's something not only cheaper, but more fun and fundamental about making it yourself. And dangerous too – but hence fun.

Most of the lab-based DIY I encounter is elementary. It comes down to using rubber bands to hold pieces of glassware together (including pieces of glassware being heated, and including post-eruption, which might have warned the chemist, but didn't), and using household appliances or beauty products to do the job of industrial equipment. Some of this is just basic, safe hardware. In my current lab, for example, we have a gun cabinet. In it we keep a Black & Decker heat gun, because the specialist lab ones are too expensive. Not to be outdone, a previous lab I worked in used a hairdrier, and kept in the glovebox a pair of hair straighteners, which were used for sealing aluminium packets by melting the seal down.

Various science teachers and science communicators have gone even more freelance. From homemade squealing jelly baby experiments (once banned because jelly babies are symbolic of humans and this imagery might induce children to torture), erupting volcanoes and hydrogen rockets made of plastic tubs, to ingesting helium, using basic salts to set soya milk, or employing hydrogen gas to set your hand on fire. In the end, they seem risky at the time, but it's not a big risk. It's a domestic level hazard and for someone who's seen her mother accidentally boil the metal on the bottom of an empty pan, it doesn't seem that large. But what do you think? To DIY or not to DIY?

Reproducibility has always been a struggle in experimental papers; we chemists moan about it, but it looks as though medicine may have exposed the underside of the ice berg.

According to Researchers at Bayer's labs, two thirds of papers are just not reproducible. And this is papers which are accepted by the scientific community as valid and quality works. In attempting to replicate 67 of them, Bayer only succeeded fully in replicating 14, and partially replicating a further 8. Naturally, this is not all due to sloppy academia: sometimes these problems are due to the complexity of biological systems, unrecognised flaws or systematic errors - very sensitive changes which may not have been observed or reported. In my experience, lack of experimental details is more often the real problem with reproducibility, but more and more often scientists are being put under pressure to publish one-time positive results because of the need to validate funding and secure promotions.

Some have mooted the idea that it should be the responsibility of peer reviewers to replicate results in journal submissions to check for reproducibility, but where they will get the funding for the equipment, machine time and chemicals for – let alone how they will create the time for this – was not fully explained in the proposal. They probably won't.

I'm a little dubious however, that reproducibility every time is a must-have (though obviously ideal). Wasn't the Rutherford gold foil experiment famously conducted by picking out only certain data – those alpha particles which came back? Or charging a droplet to measure energy quanta?

Source

There was this old 1993 paper, back when I was doing my fourth year project, where they were making potassium chalcogenides by dissolving the alkali metal in ammonia. It was dangerous. Alkali metals are obviously extremely flammable, and it also involved using very volatile, very toxic condensed ammonia, kept so with a dry ice slurry and extremely low pressures. We were trying to make a product for which the potassium chalcogenide might be a possible precursor. My supervisor looked at the method, and I thought he was going to come up with some super new modern way of doing the reaction safely.

But no.

First I did the reaction with him, then he let me do it by myself, and then he started teaching other students to carry it out.

I think something exploded just under half the time. Usually it was the glass flask that would pop, streaming a fast flow of ammonia into the room, obliging you to dash over to the cylinder and rapidly shut it off before backing away to see if anything else would happen (we probably shut off the vacuum too - I don't remember though). Although it kept going wrong, we were empowered by our initial success with it and my supervisor was convinced that we were all so terrified that we would act with maximum caution... so we kept trying. Afterall, it was fairly reliable in that, although disasters were rife and yields were low, you could pretty much guarantee that if you didn't gas the lab you would get exactly the right product you were aiming for.

But I can't help but wonder what inspired this synthesis the first time round. Did they make a habit of melting alkali metals in liquid ammonia and oohing at the pretty gold colour coming out in the deep blue liquid? Did they throw things in to this witches' cauldron and experiment with what came out as the natural course of their research or were they so determined to make potassium chalcogenides that they were driven to extreme set ups? Because if so, why did they never comment on the safety aspects of it?

We broke so much glassware. And the bigger the experiment, the more likely it was to go wrong, and the more dangerous it was when it did.

On one occasion, my supervisor was in the firing line. I wasn't working with him on this occasion - I was in the next lab down, so the first I knew of it was a shout, running footsteps and, by the time I poked my head out of the door, a host of worried faces and a discarded lab coat lying in the corridor. A few questions were enough to extract the relevant details. He had been leaning over the experiment when it went off, and alkali metal got squirted all over him. This was enough to break his blaise risk attitude - he yelled, ran out of the lab and down the corridor, ripping his lab coat from his body whilst yelling, "AM I ON FIRE?"

He was not on fire. And so the next day, we resumed the experiment.