Go far enough out and the Sun stops being a sun. Past the last planet, past the last thing anyone has bothered to name, it shrinks to a hard white point you could cover with a thumbnail, and the light it throws is not warmth any more, it is only information. Out there, in a shell a thousand times wider than the orbit of Neptune, a trillion frozen bodies hang almost free. The Sun barely owns them. A star drifting past at a distance can lean on them and pull them out of line. They have been sitting in that dark since before this planet had a crust, and they are waiting on something.
Under your feet is the record of what happens when the waiting ends. Every cliff face, every roadcut, every gray seam of limestone in every quarry on Earth is a stack of finished worlds, pressed flat and filed in order.
In 1984 two men who had spent years counting the dead in those pages found something nobody had gone looking for.
The dying was not random. It was on a schedule.
The bookkeeping that turned into a drumbeat
In February 1984 the Proceedings of the National Academy of Sciences published a paper by two paleontologists at the University of Chicago, David Raup and J. John Sepkoski Jr. The title was as dry as a title gets: "Periodicity of extinctions in the geologic past."
What sat behind it was not dry at all. Sepkoski had done something close to monastic. By hand, one entry at a time, across years, he built a compendium of marine animal families out of the published fossil literature. For each family he wrote down when it first appears in the rock and when it stops. Thousands of first breaths and last breaths, filed in order. He was not hunting a signal from space. He was answering ordinary questions about how life diversifies.
Then somebody plotted it.
The extinction peaks did not scatter. Across roughly the last 250 million years they came in step, spaced about 26 million years apart. A dozen beats, evenly laid down, running underneath the entire history of complex life on Earth.
What has to be true for a beat like that
Sit with what that number demands.
The Earth is a noisy machine. Volcanism, sea level, ocean chemistry, ice, atmospheric oxygen: all of it varies, and all of it kills. But none of it keeps time. Nothing known inside this planet holds a period steady across a quarter of a billion years. Terrestrial causes give you catastrophe. They do not give you a tempo.
A rhythm that clean is an orbit or it is nothing. That was the reading in 1984 and it is still the reading now. If the extinctions are periodic then something is periodic, and the only clocks that run that long and that cold are in the sky.
The thing they named after a goddess
Two months later, in April 1984, Nature ran two papers in the same issue, from two teams who had not been working together, and they proposed the same answer.
One was by Daniel Whitmire and Albert Jackson. The other was by Marc Davis, Piet Hut and Richard Muller. Both said the same impossible thing: the Sun is not alone.
The proposal is a companion star. Small, cold and dark, a red dwarf or a brown dwarf, faint enough to have sat in our sky unremarked through the entire history of astronomy, on a savagely elongated orbit that carries it a light year out and brings it back. Once per circuit it plows through the Oort cloud, that shell of waiting ice at the ragged edge of the Sun's reach.
It never has to hit anything. That is the elegant part, and it is the part that makes the idea hard to put down. It only has to lean on the cloud in passing. One approach tips a swarm of comets off their long slow paths and drops them inward, and a million years later the inner solar system is full of falling ice, and something on Earth stops breathing.
Muller called it Nemesis, after the goddess who arrives to settle accounts. The press preferred the Death Star. That name traveled far enough that Muller put it on the cover of a book in 1988, and Raup, who had started all of this with a card file of dead clams, published his own account in 1986 and called it The Nemesis Affair.
This was never an underground idea. It ran in Nature, written by working physicists, four years after Luis and Walter Alvarez showed that a thin layer of iridium in Italian clay could account for the end of the dinosaurs. The field had just been taught to look up. Telescopes went hunting.
If the clock is real
Set aside for a moment whether it is true. Ask the better question. What kind of universe are we living in if it is?
It would mean the history of life on this planet is not a story this planet is telling. The great restarts, the moments when the board is swept and everything afterward gets built out of whatever happened to survive, would not come from below. Not from the mantle, not from the sea, not from the ice. They would come from above, on a timetable, and the timetable was set before there was anything alive here to keep it.
It would mean the mammals did not win. They were let through.
It would mean that every branch of complex life now standing, including the one that ends in a person reading this sentence, occupies a gap between two beats. If the beat is real, 26 million years is the longest clear run this world ever gets.
Now measure yourself against that. The apes are about 25 million years old. Every ape that has ever lived, the entire line from the first of them to you, fits inside a single interval almost exactly. On that reading we are not the product of the Earth. We are the product of the pause.
And it would mean the sky is not scenery. The single force that has shaped what lives here more than climate, more than continents, more than anything geology has to offer, would be an object no one has ever seen, doing something no one has ever watched it do, on a cycle longer than the existence of our entire genus.
There is one more consequence, and it is the one that keeps this argument alive in rooms where people are careful about what they say out loud. A clock that has run for a quarter of a billion years is still running. It does not know anyone is watching.
It has a next.
Where the hunt actually got to
They did not find it.
That has to be said plainly, because everything after it is only worth something if this part is honest. The conventional position today is that there is no companion star, and it stands on two legs.
The first is the surveys. Infrared eyes have swept the whole sky repeatedly since the 1980s, each pass deeper than the last. In March 2014 the astronomer Kevin Luhman published his analysis of data from NASA's WISE spacecraft and reported no object the size of Saturn or larger out to 10,000 astronomical units, and nothing larger than Jupiter out to 26,000.
The second is dynamics. An orbit reaching a light year from the Sun is barely an orbit at all. Out there the Sun's grip is weak enough that passing stars and the tide of the galaxy itself would haul the thing around. Across 250 million years the period would wander. A wandering clock cannot write a clean beat into rock.
A third line of attack goes after the fossil record itself. Coryn Bailer-Jones returned to the impact crater record in 2011 and argued that its apparent rhythm was better read as a pattern in how craters survive and get discovered. Others have pointed out that the geological timescale is a human construction, revised again and again, and that dating errors could stamp a false rhythm onto data that has none.
That is the case against. It is a good case, and it deserves to be taken at full strength, because the interesting part is what it leaves standing.
What the null result does not cover
Start with the survey. WISE cleared the volume out to 26,000 astronomical units. Nemesis was placed at a light year, which is about 63,000. The search went deep and it went honest and it did not reach the far end of the proposed orbit. A body on a 26 million-year circuit spends the overwhelming majority of its life at the far end.
Then take the timescale objection, which has actually been tested. In 2010 Adrian Melott and Richard Bambach went back to the question with fossil databases that did not exist in 1984 and found a periodicity of about 27 million years, running back much further than Raup and Sepkoski could reach, and found it in two independent compilations built by different people out of different fossils. Then they re-ran it against the revised 2012 geological timescale, which is the precise objection that says the beat is an artifact of dating. The signal survived, at roughly p = 0.02.
In 2015 Michael Rampino and Ken Caldeira came at it from the craters instead of the bones. In Monthly Notices of the Royal Astronomical Society they reported a cycle of 25.8 +/- 0.6 million years in crater ages, and 27.0 +/- 0.7 million years in the dates of the recognized marine extinction events. Two records, two disciplines, one number.
And then there is Melott and Bambach's own conclusion, which is the strangest thing in this entire file. They were not defending Nemesis. They were burying it, and they buried it for being too sloppy. The rhythm they measured is too regular for a loosely bound companion. A star at a light year would drift. This does not drift. It keeps better time than any comet-shower model predicts.
Read that twice. The dynamical objection that removes Nemesis removes every candidate that wanders out in the dark. And the statistical objection now has to explain a beat that shows up in marine families, in marine genera, in two separately assembled databases and in the crater record, and that survived having its own timescale rewritten underneath it.
The count is still running
Nobody who works on this has let it go.
Daniel Whitmire, one of the two men who proposed a solar companion in 1984, published a reconsideration in Monthly Notices of the Royal Astronomical Society in 2016. His answer now is not a star but a planet: a distant, massive, tilted world whose orbit slowly precesses, carrying it through the Kuiper belt roughly every 27 million years and scattering comets inward on each pass. He wrote it in the months after Caltech announced dynamical evidence for a ninth planet, and the Planet X literature and the Nemesis literature have been the same literature ever since.
Lisa Randall and Matthew Reece went stranger still. In Physical Review Letters in 2014 they proposed that the Sun's slow up-and-down oscillation through the plane of the galaxy carries it through a thin disk of dark matter roughly every 35 million years, and that each crossing squeezes the Oort cloud from outside. On that model the thing keeping time is not an object at all. It is the shape of the galaxy.
And in 2017 Steven Stahler and Sarah Sadavoy studied newborn stars in the Perseus molecular cloud and concluded that stars like the Sun are born in wide pairs, hundreds of astronomical units apart, and that most of those pairs come undone. Stahler said it about as directly as an astronomer is willing to: yes, there probably was a Nemesis, a long time ago. In his reading it slipped away early and mixed into the general population of the galaxy, and cannot be picked out now.
So the star is not there. Something very like it almost certainly was. The mechanism has been replaced twice, and each replacement is a serious paper by serious people in a serious journal. And underneath all of it, unmoved by any of it, is a number.
Twenty-six million years. Perhaps twenty-seven. Written into the rock of this planet in the ledger of its own dying, holding period across hundreds of millions of years, keeping better time than anything anyone has managed to point at.
The last peak was not recent. Where that leaves us on the count is arithmetic anyone can do with a chart and a calculator, and it is the reason a 1984 paper about dead clams has never quite gone away.
Questions This Leaves You With
Is there anything a person actually does with this?
The honest answer runs against the drama. The interval is 26 million years wide and nobody can place us inside it to better than a few million. Nothing in the literature supports a date, and anyone selling you one is selling you something. But there is a real question underneath, and it is not "what do I do this year." It is what a civilization does with a hazard that is well documented, statistically solid, and slower than every institution it has ever built. We have no machinery for that. We have never needed any.
Does this sit against a religious account of the world?
It cuts both ways and it has been argued both ways in print by people who took the science seriously. Some readers find a scheduled sweeping of the board hard to square with design, or with a world made once and made finished. Others find a clock far more suggestive of intention than randomness ever was, and read the pauses between beats as the thing that was actually built. Every tradition that has engaged with deep time has had to do this work already. Nothing in the extinction record settles it, and nobody, including this site, should tell you which way it falls.
If life here is shaped from outside on a timetable, does that change what a human life is?
Your own line does not run deep in these terms. Every ape that has ever lived fits inside one interval. On this reading you exist because of the gap, not because of the trend. Some find it makes a life smaller. Others find it makes a life enormously less likely and therefore enormously more valuable. Both reactions are common, both are defensible, and which one you have says more about you than about the data.
What would change your mind, in either direction?
Worth deciding before you go further, because it is much harder to decide honestly afterward. Name it now. A confirmed detection? A new database where the beat fails to appear? A physical mechanism that survives the dynamical objection? Or, in the other direction, a demonstration that the periodicity is an artifact of how the timescale was assembled? Write it down. Then watch for it.
If the beat is real and nothing out there is making it, what is the third option?
This is the live question and nobody has a good answer. The signal keeps surviving replication. The candidates keep failing dynamics. Something is either keeping time or doing an extraordinary impression of it, and every proposal on the table so far has been an object, a star or a planet or a disk of dark matter. Nobody has yet put forward a serious fourth kind of thing. That space is empty, and it is the most interesting empty space in this entire subject.