A mayfly hatches at eleven at night.
By the time it can think at all, it already knows one thing with total certainty: it was born into darkness, and darkness is all there has ever been. If it could reason, it would reason correctly, from everything it has ever measured. It would build a theology around the dark. It would explain the cold. It would, if the night stretched on long enough, begin to notice small rhythms within the blackness — and it would be right about those too. What it cannot know, because nothing in its lived experience contains it, is that the sun is already on its way, and was never in question, and has nothing to do with anything the mayfly did or believed or feared in the dark.
The mayfly is not wrong. It is short.
Every rhythm that has ever been proven real was proven by a witness who outlived it. That sentence sounds obvious until you ask how rare the condition actually is.
In the early sixteen-hundreds, Francis Bacon wrote — from observation alone, no instruments, nothing but a long enough life and a habit of writing things down — that English weather seemed to repeat itself “in every five and thirty years.” He had no way to confirm it. He simply noticed it once, and again, and wrote it before he forgot.
Two hundred and sixty years later, in 1891, an Austrian meteorologist named Eduard Brückner isolated the same rhythm from instrument records Bacon never had access to — a roughly thirty-five-year cycle in weather and glacial advance, found independently, with no citation of Bacon at all. Two witnesses, separated by more than two centuries, arrived at the same number without comparing notes, because the rhythm was there whether or not either man happened to be watching.
The Carnegie Institution’s sunspot records, kept from 1835 to 1930, show something similar: an 11.4-year rhythm in solar activity, and — measured independently, on the ground — an almost identical rhythm in terrestrial magnetism. Three separate research groups working weather data in the early twentieth century — Brunt, Gillette, Anderson — each landed on a rhythm close to 22 years, without setting out to confirm one another. A 41-month rhythm measured in the solar constant by the Smithsonian shows up again, same length, in atmospheric electricity readings taken across the ocean in London.
None of this proves what causes what. The men who compiled it said so themselves, in print, in 1947 — warning their own readers against exactly the temptation to skip from correlation to causation. What it proves is narrower and stranger: that independent witnesses, decades or centuries apart, keep finding the same numbers in nature, whether or not anyone was looking for them to agree.
There is a piece of arithmetic buried in an appendix of that same 1947 study that deserves to be better known than the study itself. It concerns the moving average — the standard tool used to smooth noisy data into a clean, readable trend.
The math says this: a moving average shorter than a rhythm’s true period will weaken that rhythm on the chart. A moving average equal to the rhythm’s period will erase it completely. Not obscure it — remove it, mathematically, from the picture, as if it had never been there.
This is not a metaphor. It is what a moving average does, by definition, to any cycle exactly as long as the window used to calculate it. The tool built to reveal a trend is, under specific and unremarkable conditions, also a tool that can make a real rhythm vanish without anyone touching the underlying data or committing anything that looks like fraud.
The research continues every week.
Hidden history. Forgotten symbols. New articles before they disappear into the archive. Delivered once a week. No noise.
Join the dispatch →Prefer Telegram? Daily updates here.
Which raises the only question this piece is actually asking: when a modern chart shows no long rhythm — is that because there isn’t one, or because someone, knowingly or not, chose a window exactly wide enough to erase it?
Step back further than any human record goes, and the witnessing problem gets worse before it gets better.
The full instrumental temperature record — thermometers, consistent methodology, global coverage — runs a little under two hundred years. That is barely five turns of the Brückner cycle laid end to end. It is a single human lifetime and a half.
Ice cores pulled from Antarctica hold a different kind of record. The EPICA Dome C core preserves a continuous climate signal reaching back roughly 800,000 years. Read against it, the planet does not hold one climate — it swings, on a nested set of rhythms driven by the shape of Earth’s orbit and the tilt of its axis: roughly 100,000 years, roughly 41,000 years, roughly 23,000 years, stacked one inside the other, the way a shorter wave rides a longer one. Milutin Milanković worked these numbers out by hand, from orbital mechanics, before satellites existed to confirm them — a witness, again, patient enough to calculate what he could not live to see proven.
Inside that longer record: the Eemian interglacial, roughly 120,000 years ago, ran warmer than the present in significant respects. The Holocene Climatic Optimum, a few thousand years back, ran warmer than today across large parts of the world. Further back still, eras existed with no ice at either pole at all. None of this is fringe material — it sits in the same paleoclimate archive every institution cites. It simply rarely gets placed, in the same sentence, next to a two-hundred-year instrumental window and asked to explain what that window is actually wide enough to see.
Not every long rhythm stays out of human reach forever — some have been caught, but only by witnesses who understood that no single lifetime would be long enough to catch them alone.
Venus returns to the same position against the sun and Earth every 583.92 days. Repeat that five times, and 2,919.6 days pass — within a few hours of exactly eight years. Plotted against the stars across those eight years, Venus does not wander. It traces a five-pointed rose, closing almost perfectly on itself: the pentagram, not as symbol first, but as the literal shape of an eight-year orbital record.
Babylonian astronomers recorded this cycle with enough precision that the surviving Venus Tablet of Ammisaduqa preserves observations reaching back more than three thousand years. On the other side of the world, with no contact with Babylon, Maya astronomers tracked the identical 584-day cycle to comparable precision in what survives as the Dresden Codex. Neither tradition caught the pattern in one lifetime. They caught it because one generation of watchers wrote the number down faithfully enough for the next generation to pick it back up exactly where it had been left.
That is the only known cure for the mayfly’s problem: not a longer life, which no witness gets to choose, but a record passed forward with enough discipline that the truth doesn’t depend on any single witness surviving to see it confirmed.
The sun was never in question. It was only ever a matter of how long the mayfly had been asked to wait, and whether anyone had bothered to tell it there was a record.
