Shower curtain drift
A Socratic walk-through of shower curtain drift — reasoned out one step at a time, not lectured.
The question we started with
THE QUESTION #Why does a shower curtain swing inward against your legs instead of being pushed away by the water?
A shower fires water at the curtain. Water carries momentum. So the curtain should be pushed outward, away from you — and instead it billows in and wraps itself around your shins, which is the one outcome the obvious force predicts against.
There is a temptation to reach straight for a tidy answer. I want to resist it, because this is a question the physics community has genuinely not closed, and watching why it stays open teaches more than any of the candidate answers does.
Reasoning it through
REASONING #Begin with the one thing everyone agrees on. A curtain is a sheet of thin plastic hanging free. To push it inward you need a pressure inside the enclosure slightly lower than the pressure in the bathroom. There is no other handle on a limp sheet.
So how small is "slightly"? A pressure difference of even a fraction of a pascal across a curtain's area is enough to move it, against an atmospheric pressure of about a hundred thousand pascals — a disturbance on the order of a millionth of the ambient. That is the crux of the puzzle: the effect is so easily produced that almost any plausible mechanism is strong enough. Sufficiency proves nothing. The hard question is not what could do it, but what does.
So let us put the candidates in a row and ask what each predicts.
The first is entrainment. The spray drags air along with it — droplets transfer momentum to the air they fall through — so a downward current runs inside the enclosure. Fast-moving air within a flow has lower static pressure than still air alongside it, and something loosely Bernoulli-shaped follows. The trouble is quantitative: the speeds involved are modest, and it is not obvious the resulting drop is where the curtain is rather than where the droplets are.
The second is buoyancy — the chimney account. Hot water warms the air, warm air rises and escapes over the top of the rail, and air must be drawn in low to replace it, dragging the curtain's lower edge inward on the way. This one has a satisfying shape and it explains why the bottom of the curtain moves most. It also makes a prediction we can test in the shower tomorrow: it should not happen with cold water. But it does happen with cold water. Reduced, perhaps, but there.
The third is the one that became famous. In 2001 David Schmidt, an engineer at the University of Massachusetts Amherst, ran a computational fluid dynamics simulation of a shower and found the flow organising itself into a vortex with a horizontal axis — a sideways whirl filling the enclosure, its low-pressure core sitting near the curtain and sucking it in. It is an elegant result, and it earned an Ig Nobel Prize, an award for work that makes you laugh and then think. It also has a virtue the buoyancy story lacks: a vortex driven by the spray's momentum does not care about temperature.
So is that the answer? Here is where honesty has to win. Schmidt's result is a simulation of a particular idealised shower, and the strong verification you would want — measuring the pressure field and the circulation in a real bathroom and matching them — has not, to my knowledge, been convincingly done. Meanwhile the effect varies with the geometry of the stall, the shape and pressure of the head, whether a window is open, whether the door is shut. It is entirely possible that all three mechanisms operate, in proportions that shift from bathroom to bathroom.
The analogy
THE ANALOGY #Think of a detective with three suspects, each of whom had the means, the motive, and no alibi. The crime required so little strength that being capable of it eliminates nobody. The case cannot be closed by asking who could have done it; only by finding evidence that discriminates — a witness who saw the deed, or a fact one suspect cannot account for.
A crime has one culprit, whereas fluid mechanisms superpose freely — all three "suspects" may be acting at once, and the honest verdict may be a shifting mixture rather than a name.
Clarifying the model
THE MODEL #The instinct that the spray should push the curtain outward is not wrong, only outweighed. The water does deliver outward momentum to whatever it strikes — but most of the spray lands on you and the floor, not the curtain, while the air motion it sets up acts over the curtain's entire area at once. A weak pressure field distributed everywhere beats a stronger force concentrated where the curtain mostly is not.
It is also worth being clear about what "Bernoulli" is and is not doing. The principle relates speed and pressure along a streamline in steady, low-loss flow; a bathroom full of turbulent, humid, droplet-laden air is not that, so invoking it by name gestures at a mechanism rather than deriving one. The same caution applies to the vortex account, which is a claim about what a simulated flow field organises itself into, not a closed-form result.
None of which stops the fix from working. Weighted hems, magnets, a curved rail, or simply tucking the curtain into the tub all defeat the effect, because they oppose a force that was only ever a fraction of a pascal.
A picture of it
THE PICTURE #How to readThe three boxes labelled O1 to O3 are observations any explanation has to survive, not claims about a mechanism. The three elements below are the candidate accounts, and an arrow means that account comfortably delivers that observation. Read the missing arrows, which carry the argument: the buoyant-chimney story has no link to O1 because a cold shower ought to switch it off and does not, while entrainment has none to O2. The vortex account reaches all three, which is why it is the best known — and O3 is the reason none of this settles the matter, since so little force is required that being sufficient makes no candidate the culprit.
What became clearer
WHAT CLEARED #The curtain moves inward because the air inside the enclosure sits at very slightly lower pressure than the bathroom — and that "very slightly" is the whole difficulty. The margin is so small that entrainment, buoyancy and a spray-driven horizontal vortex are each individually adequate, so the question cannot be settled by showing that a mechanism works. The best-known account is Schmidt's vortex, which alone survives the cold-shower test, but it rests on simulation rather than measurement, and the mixture probably differs from one bathroom to the next. Some entirely ordinary things are still open.
Where to go next
ONWARD #- How you would actually measure a sub-pascal pressure field in a wet, turbulent bathroom.
- What other domestic puzzles — the Brazil nut effect, the cracking of a whip — took decades to settle.
Key terms
TERMS #| Term | What it means |
|---|---|
| Entrainment | the dragging of surrounding air into motion by a stream of falling droplets. |
| Bernoulli's principle | the relation between higher flow speed and lower static pressure along a |
| Horizontal-axis vortex | a rotating body of air whose axis lies flat, as proposed by Schmidt's |
| Pascal | the SI unit of pressure; the atmosphere is around a hundred thousand of them, and the |
Every term the collection defines is gathered in the glossary.