I didn't know until this week that Joule was doing his best work while the family brewery was still going. Beer and thermometers, more or less the same trade: if you can't measure a degree properly you can't make the same ale twice. So he already had instruments finer than anything I ever had in a surgery, and a habit of trusting them.
The experiment itself is kitchen-simple. A tub of water with a paddle in it, insulated. A weight on a string, falling slowly, turning the paddle. You know how much the weight weighs and how far it drops, which gives you the work. Then you read the thermometer. The water warms — barely, a fraction of a degree, the sort of change you'd never feel with a hand in the tub. Stir your tea hard enough for long enough and you are doing the same thing, just too feebly to notice. Work going in, heat coming out, and a number connecting the two.
What I admire is the patience of it. Run it, let everything settle, correct for the heat leaking out of the tub, run it again. He was doing this from about 1843 onward and kept refining it for years. He died at Sale in Cheshire in 1889, by which point the idea had stopped being his and started being simply how things are. There's a portrait of him late on — white beard, dark coat — looking rather like every careful man I've ever worked with.
Ada, the brewery is the whole reason those thermometers were so good — mash temperature is a few degrees' worth of difference between a sweet wort and a thin one, so brewers were fussing over fractions long before anyone called it physics. He grew up with instruments that had to be right or the ale came out wrong.
Ada, the lovely bit is that the joule doesn't mention water at all now — it's kilogram metre squared per second squared, and since the 2019 redefinition the kilogram itself hangs off the Planck constant. One joule is roughly lifting an apple a metre, which is a long way from a tub and a paddle.