In 1770 the Opéra Royal at Versailles burned some three thousand candles a performance, and somebody had to walk the footlights and cut them back. In German-speaking theatres the job had a name — Schnäuzer — and often a jester’s costume, since the man worked in full view of the house. A Swiss National Museum history puts the routine at trimming each wick to ten or fifteen millimetres, every five to twenty minutes. Miss a round and the flame guttered and smoked.
The reason is the whole problem. Old wicks did not burn. The wax burned; the string only delivered it, so as the candle sank the wick stayed and the exposed length grew. A taper is a foot tall and burns about fourteen hours: with a wick that will not consume itself, you finish the evening holding a foot of char with a sputter on top.
Now measure a candle you can buy today.
In 2005 Anthony Hamins and Matthew Bundy at NIST, with Scott Dillon of the ATF fire lab, set 21 mm paraffin tapers on a load cell in a draught-proof box and watched for two hours. The candle loses height steadily: 0.35 millimetres a minute, dead linear. The flame does not. After fifteen minutes it settles at forty-two millimetres, give or take one, and holds there for the remaining hundred and fifteen while forty millimetres of candle vanishes underneath it.
Forty millimetres is the height of the flame: unchecked, the wick would end that run as tall as the fire it feeds. Instead it holds near twelve — the lowest four below the flame’s base, never burning at all. Something cuts it back at 0.35 millimetres a minute, matched to the wax, all night, to a tolerance no one with scissors ever managed.
The something is a twist. A modern wick is a flat braid with one side under slightly more tension than the other, so as it chars it bends — roughly the way a ribbon curls dragged across a scissor blade. NIST had to put the lean into their model to make the numbers work: a strand twelve millimetres tall, leaning five off the centreline. They could see it in the heat, too — the flux above the flame runs higher on the side the wick points, because the flame tip sits over the curl, not the candle.
Here is where I had it backwards — and so does most of what you will read.
The tempting story is that the curl carries the tip out of the hottest part of the flame into cool air, where it burns off safely. I said something close to that at the end of the last issue. It is the wrong way round. Faraday showed as much in 1860 by holding a piece of paper flat against a flame: the paper chars in a ring. “Do you not see that it is not in the inside?” he asked. The middle of a candle flame is not the hot part. It is cool fuel vapour with no air in it — “in the middle of the flame, where the wick is, there is this combustible vapor; on the outside of the flame is the air.”
A wick standing straight up sits in fuel with no oxygen, and cannot burn however hot it gets. The curl is not rescuing the tip from the fire. It is delivering it, out through the reaction sheet to the one place in the flame with air enough to destroy it. Heat was never the operative variable.
Which makes the loop legible, once you stop looking for a thermostat. Too much wick and more tip pushes past the ring into open air, and burns faster. Too little and the tip withdraws into the airless core, stops burning, and waits while the receding wax uncovers more. The set point is stored nowhere: it is a distance — wax pool to reaction sheet — fixed by the flame, which is fixed by how much wick delivers fuel. Regulator and regulated are one object, running on under a tenth of a percent of the candle’s mass.
One honest gap. The curl is measured, but why an asymmetric braid bends as it chars does not appear tested in any combustion journal I could find. Manufacturers say differential tension; so do the museums. Take the geometry as data and the reason as reported.
Braided wicks arrive in the 1820s. Goethe had wanted one badly, writing to Charlotte von Stein in 1779: “I cannot think of a better invention than lights that keep burning without needing to be trimmed.” By 1860 Faraday held one up as unremarkable — “a plaited wick, which does not require snuffing” — and plainly thought it the loveliest arrangement in the room: “I cannot imagine a more beautiful example than the condition of adjustment under which a candle makes one part subserve to the other to the very end of its action.”
No. 23 followed a fire in Rome that stayed lit a thousand years and found the thing standing was not a flame but a duty roster — a loop that failed the moment nobody came for the next shift. A self-trimming wick is the other outcome. The job did not get easier and it did not go away. Somebody handed it to the string, and everyone forgot there had ever been a job.
≈
One loop I’m watching
Next: what a flow does when it cannot get past an obstacle cleanly — and this time it does not stand still, it sheds. Downstream of a blunt body the vortices peel off one side, then the other, and travel away. The street they make does not travel. It sits there, evenly spaced, for as long as the flow stays inside a band of speeds narrower than you would guess. Too slow and there is nothing to see; too fast and it tears up into noise. Next time.