Field notes on things that run themselves
Squeeze It and It Gets Darker
δ Cephei is a yellow supergiant, and it will not hold still. It swells and shrinks on a period of five days, eight hours, thirty-seven minutes. Noticed in 1784, it has kept that beat ever since. Nothing outside the star sets it: no companion, no orbit, no arriving signal. The period is a property of the star’s own body, the way a note is a property of a pipe’s length. The surprise isn’t that a star can ring; it’s that this one doesn’t stop.
Arthur Eddington put the difficulty plainly in 1917. Any oscillation leaks energy; something must pay it back. For that, he wrote, the star “must behave as an engine”: taking in heat when hotter than average, giving it out when cooler — “just the opposite of what usually happens in natural conditions.” He had the analogy right: radiation as steam, a layer of the star as piston, somewhere a valve. He also had the honesty to add that how this comes about “must be left unsolved.”
It stayed unsolved for decades, because stellar material does the opposite. Opacity — how badly a gas obstructs radiation — falls roughly as density divided by temperature to the three-and-a-half power. Squeeze ordinary stellar gas and both climb, but temperature climbs faster, so the parcel goes more transparent. Heat escapes it fastest exactly when it is being pushed. Ordinary stars are shock absorbers, damping their own wobbles — which is why the sky isn’t full of throbbing stars.
Unless the temperature refuses to climb. Bury a layer where helium is halfway through losing its second electron, and compression energy has somewhere else to go: it strips electrons off more atoms instead of heating the gas. Temperature barely moves. Density rises anyway. So opacity rises: the layer gets darker as it is squeezed, dams the radiation below it, and pressure builds until the envelope above is shoved out. Expanding, the ions take their electrons back and hand the energy over; opacity falls, the dam opens, the heat floods out, the cooled layer drops. Once per cycle, as long as that layer stays put. The opacity swing is the κ-mechanism; the same ionisation’s pull on neighbouring heat, the γ-mechanism.
Finding it took another thirty-five years, and the first guess was wrong. Eddington’s 1926 book offered two engines: one modulating the star’s nuclear burning under compression, one needing no such help. The nuclear version was favoured, because nobody yet knew how tightly a giant packs its mass. Around 1950 Schwarzschild’s Princeton group used ENIAC to model giants properly, and Epstein drew the consequence: displacement in the burning core is about a millionth of that at the surface. The core barely moves; whatever is pumping cannot be there.
Sergei Zhevakin found the layer in 1953 — in Russian, at the height of the Cold War, barely read in the West until an Annual Review invitation carried it over a decade later. John Cox got there sideways, explaining something else entirely: a Cepheid is brightest slightly after it is smallest. Models settled it by 1962, with a pleasing symmetry: of the star’s two partial-ionisation zones, the deeper helium one drives the pulsing, while the shallower hydrogen one, routinely miscredited as the engine, causes the very lag that led Cox to look.
It also explains the rarity. The valve has to sit at just the right depth: too shallow and there is too little mass above it to push, too deep and convection ferries the heat around it. That narrow allowance is the instability strip; why some stars inside it stay quiet is still open.
Then the part nobody predicted. Because a Cepheid’s period is set by its own structure, it reports on that structure. Henrietta Leavitt, working at Harvard from glass plates at thirty cents an hour, measured variables in the Small Magellanic Cloud — all at one rough distance, so their brightnesses compared directly — and found the slower ones are the brighter ones — turning a period into a brightness, and a brightness into a distance. Twenty-five stars, published in 1912 under Edward Pickering’s byline as convention demanded, though he credited her fully in his first paragraph. A 2025 reanalysis returned to her notebooks: modern methods halve the scatter, and her result agrees excellently with today’s. That ladder still reaches, and the quarrel between its top rung and the early universe’s accounting is among the liveliest in physics.
A star in balance holds still because pressure and gravity have agreed on a number (No. 6). This one never quite agrees, and the disagreement has a period. It isn’t a clock the star consults, as a cell keeps a schedule light re-phases without powering (No. 67), nor the quotient of two competing transports across a fluid (No. 61). It is the round-trip time of the energy itself through one layer that won’t pass it at a steady rate. There is no timekeeper. There is a delay, and the delay is the clock.
One loop I’m watching
Next: a bird’s egg looks sealed, and it isn’t. The shell is a mineral crust drilled through with thousands of microscopic pores, and for about three weeks oxygen walks in through them while carbon dioxide and water vapour walk out — pulled by nothing but the gradient the embryo’s own metabolism creates. No pump, no muscle, nothing tending it. A still object stays alive by having got its own leakiness exactly right, and then breaks the shell it was breathing through.
Tip: the ← and → arrow keys move between issues.
Prefer feeds? Follow by RSS or JSON Feed.
Want something playable between issues? Thought Toys has its own email list and RSS feed.
New to The Standing Wave? Start here →