Field notes on things that run themselves
Always Ten Days Old
Behind this sentence, better than a hundred million cells in each of your eyes are throwing away the part of themselves that does the seeing. Every rod and cone ends in an outer segment: a stack of roughly a thousand thin membrane discs, each packed with the pigment that catches light and starts the signal. The stack isn’t permanent. Fresh discs are assembled at the bottom and the column pushed slowly upward; at the top the oldest are pinched off in packets of eight to thirty and swallowed whole by a different cell waiting behind. A rod builds eighty to ninety new discs a day. Within nine to thirteen days — depending on the animal and who’s counting — none of the original stack is left. The cell itself can last a lifetime; its seeing end is never more than a week or two old.
We know this from a beautifully plain experiment. In 1967 Richard Young injected rats, mice and frogs with radioactive methionine, an ordinary amino acid the cells duly built into new protein. He took retinas on successive days afterward and laid them against photographic emulsion. What showed up was a band: a narrow stripe of radioactivity near each rod’s outer-segment base, marking exactly the discs made while the label was available. Days later the band sat higher. Later still, higher again. Eventually it reached the tip and vanished. Nothing had grown or shrunk. The band was a cohort of discs riding up through a structure of constant length, like a chalk mark on a conveyor belt, and falling off the end.
The disposal isn’t a steady trickle either. It comes in a daily burst, on a clock the eye keeps itself. In rats on a light-dark cycle, Matthew LaVail found in 1976 that shedding spikes shortly after the lights come on: swallowed disc packets in the cell behind run two and a half to five times more numerous than at any other hour. The obvious reading is that light triggers it. It doesn’t. Held in continuous darkness, those retinas went on producing the same burst at the same subjective hour for at least twelve days running, and the interval settled slightly long: an estimated twenty-four hours and eight to nineteen minutes. That overshoot is the signature of a real internal oscillator running free, the same signature the body’s master clock shows (No. 25). Dawn doesn’t cause the shedding; it corrects it — slowly. A three-hour shift in the lighting takes the rhythm at least four weeks to absorb.
Young’s method had a blind spot, and it took a decade to find. Repeating it in rhesus monkeys in 1971, he watched the rods band as expected while the cones showed only a weak, diffuse haze — no cohort, no stripe — and concluded they must renew some other way, swapping molecules in place rather than whole discs. The structure is why: a rod’s discs are sealed separate packets, so a pulse of label locks into one legible batch, while a cone’s are a continuous ruffle of infolded surface membrane, open to the outside, through which new protein simply spreads. By 1977 Steinberg and colleagues could watch the neighbouring cell reach out and engulf cone tips directly. Cones shed too; the technique that revealed the answer in rods had been incapable of showing it in them. Their timing differs untidily besides — rod shedding peaks after light onset, cone shedding in several species falls in darkness, and in cats both track light onset.
The swallower is the retinal pigment epithelium, a single dark layer directly behind the photoreceptors. Each of its cells caps dozens of rods and cones and eats their discarded tips daily, forever; one estimate puts a single cell’s lifetime total past a billion discs across seventy years. In an adult eye these cells essentially never divide — the same ones do the same job the whole way through, which is exactly why the arrangement is fragile. When the disposal falters, the debris has nowhere to go.
This isn’t a cell digesting its own worn parts, the way autophagy does (No. 45): that machinery runs inside one cell and speeds up when resources get scarce. Here the waste crosses a cell boundary intact and becomes a neighbour’s problem, on a schedule that doesn’t negotiate. It isn’t the orb-weaver’s overnight rebuild either (No. 39) — a web comes all the way down, and there’s an hour when no web exists. An outer segment never gets that gap; it’s replaced from underneath while still working, the only way the job could be done at all. And it’s the exact inverse of last issue’s fallen pine cone (No. 66), dead tissue life built once and walked away from for good. This is tissue life cannot walk away from for a day.
An outer segment looks like an object. It behaves like a queue — material entering at one end, leaving at the other, the length between them holding steady only because the two rates match. The stack never ages, because it is never the same stack. It just stays about ten days old.
One loop I’m watching
Next: a Cepheid variable star doesn’t shine steadily. It swells and shrinks on a period of days to weeks, and nothing outside the star sets that period — a buried layer of half-ionised helium gets more opaque as it’s squeezed, so it traps the radiation trying to leave, builds pressure, drives itself back outward, goes transparent, releases the heat and falls. A valve made of opacity, pumping itself once per cycle. And because that period turned out to be tied to the star’s true brightness, a rhythm the star makes for itself became a ruler for measuring the universe.
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 →