The impact has a date and an address. Argon-argon (40Ar/39Ar) dating puts it at 66.043 million years ago, give or take about eleven thousand years, one of the most precisely fixed events in all of deep time. The impactor struck the Yucatan Peninsula of Mexico, near the modern town of Chicxulub Puerto, at about 21.4 degrees north and 89.5 degrees west. Chromium isotopes in the boundary layer identify it as a carbonaceous chondrite asteroid, roughly ten to fifteen kilometers across.

The crater it left is about 180 kilometers wide, a buried multi-ring structure, with a transient bowl around 100 kilometers across at the moment of impact. The energy released was on the order of 4.2 x 10^23 joules, roughly 72 to 100 teratonnes of TNT, beyond any natural event in recorded human experience.

When the dust settled, roughly three-quarters of all species on Earth were extinct. The primary figure in our records is about 76 percent, though one of our own catalogs gives 75; either way it was a near-total culling of the tree of life. All non-avian dinosaurs died, a clean 100 percent, and they are only the most famous of the dead. What makes the K-Pg extinction unique among the great dyings is that it has a single, clear primary trigger, which sets it apart from murkier catastrophes like the Permian-Triassic Great Dying.

In 1980 a team at the University of California, Berkeley, the physicist and Nobel laureate Luis Alvarez, his son the geologist Walter Alvarez, and the chemists Frank Asaro and Helen Michel, reported something strange in the boundary clay at Gubbio, Italy. It was loaded with iridium, an element that is vanishingly rare in Earth's crust but common in asteroids and meteorites. The enrichment ran from thirty to a hundred and sixty times the normal background level.

That was not a local fluke. The same iridium spike was soon found at more than a hundred K-Pg boundary sites, on every continent and in deep-sea cores, a single global signature laid down at the same instant everywhere. It is now the anchor of the most complete set of impact markers in the entire geological record: the iridium, worldwide shocked quartz, altered glass microspherules, nickel-rich spinels, wildfire soot, and an osmium-isotope shift to extraterrestrial values, all in one clay layer.

The reception was frosty. Many paleontologists had built careers on gradual-extinction models and did not welcome a couple of physicists telling them a rock from space had done it in an afternoon. The argument ran for a decade. It ended not with a debate won on rhetoric but with the discovery of the murder weapon itself: the crater.

The crater had, in a sense, already been found. In 1978 the geophysicists Antonio Camargo and Glen Penfield, surveying for the Mexican oil company Pemex, spotted a vast buried circular anomaly in the gravity and magnetic data beneath the Yucatan. He suspected it was an impact structure but could not publish, because the data were proprietary. It was only in 1991 that Alan Hildebrand, at the University of Calgary, independently pieced the structure together and, publishing in the journal Geology with Penfield, tied it to the K-Pg boundary. The buried crater even has a surface trace: a ring of water-filled sinkholes, the cenotes, arcs across the Yucatan along its rim, visible from orbit.

In 2016, IODP-ICDP Expedition 364, co-led by Joanna Morgan of Imperial College London and Sean Gulick of the University of Texas, drilled 835 meters of continuous core out of the crater's peak ring, the inner ring of hills that only the largest impacts raise. The core was unambiguous: shocked minerals, melt rock, and suevite, the shattered breccia of a hypervelocity impact.

The peak ring told a stranger story still. Its rocks are granite, hauled up from deep in the crust, and the core showed they had risen through a process called acoustic fluidization. For a few minutes after impact the solid crust behaved like a fluid, rebounded upward by roughly ten kilometers, overshot the center, collapsed outward, and froze in place, exactly as Jay Melosh of Purdue University had predicted from computer models years earlier. This is what a complex crater is: not a simple bowl but a structure with a central peak or peak ring thrown up by the ground rebounding, the same class of feature seen at Manicouagan in Quebec, only far larger.

Then the core delivered the detail that ties the crater to the killing. The Yucatan bedrock is rich in gypsum and other sulfate minerals. The peak-ring core had almost none. All that sulfur had not stayed in the ground; it had been vaporized and blown into the sky, which is precisely what the impact-winter model needs in order to shut down the Sun.

Later modeling sharpened the picture of the strike itself. From the crater's asymmetric shape, Collins and colleagues concluded in 2020 that the asteroid came in at a steep angle, roughly forty-five to sixty degrees from horizontal, traveling from the northeast, close to the worst-case geometry for lofting climate-altering gas high into the atmosphere. It is a model result rather than a direct measurement, but a well-constrained one.

For scale: only a couple of hundred impact structures are confirmed on the whole planet, most long since erased by erosion and plate tectonics. Chicxulub, at 180 kilometers, is among the largest and best preserved. Only two confirmed craters clearly outrank it in size, both far older and deeply eroded: Vredefort in South Africa, about 300 kilometers across and two billion years old, and Sudbury in Ontario, originally around 250 kilometers, whose ancient wound still hosts one of the richest nickel deposits on Earth.

Two more markers close the case on how bad it got. A soot and charcoal layer sits in the boundary clay worldwide, the residue of wildfires on a scale hard to imagine. And directly above it comes the fern spike, a sudden flood of fern pollen. Ferns are the weeds of catastrophe, the first plants to recolonize scorched ground, and their pollen marks the moment the recovery began.

There may even be a piece of the asteroid itself. In 1998 Frank Kyte reported a 2.5-millimeter fossil meteorite fragment preserved in boundary clay from a Pacific deep-sea core, a possible surviving splinter of the impactor. Its identification is not universally accepted, but if it holds, it is a physical shard of the object that ended the Mesozoic.

The core preserves the sequence as cleanly as a stack of pages: impact melt rock at the bottom, then coarse sediment dumped by the returning tsunami, then the fine material that drifted down as the water stilled, all laid down in about a day. The extinction that followed was, by geological standards, instantaneous. High-precision dating brackets the whole event inside roughly thirty-three thousand years, and the killing blow itself likely took decades to centuries, not millions of years.

The primary killer, on the current reading of the evidence, was impact winter. The dust and sulfur aerosols lofted into the stratosphere blocked an estimated eighty to ninety percent of sunlight, global temperatures fell by around ten degrees Celsius, and photosynthesis collapsed. Everything else came as a cascade: seismic shaking and global landslides in the first minutes, a pulse of infrared heat as ejecta rained back through the atmosphere, megatsunamis, then months to years of cold and dark, acid rain from vaporized rock, and a battered ozone layer letting ultraviolet through. Impact winter starving the food web from the bottom up is the leading kill mechanism, not a settled certainty.

Climate modeling by Bardeen and colleagues in 2017 found that soot alone could have driven three to sixteen years of subfreezing temperatures across the mid-latitudes. Whatever the exact number, it was long enough. A food web that runs on sunlight does not survive years without it.

Chicxulub was also, in a grim sense, unlucky in where it hit. The Yucatan target rock was rich in sulfate-bearing evaporites and carbonates. Vaporizing those rocks threw far more climate-wrecking sulfur into the air than a strike on granite or basalt would have. The same asteroid landing on different ground might have been a far smaller catastrophe.

The pattern of who lived and who died fits this exactly. The hardest hit were the animals and plants tied most directly to sunlight-fed food webs: the non-avian dinosaurs, the flying pterosaurs, the marine reptiles like mosasaurs and plesiosaurs, the coiled ammonites, the reef-building rudist clams, and most of the floating plankton. The survivors were the small, the buried, and the undemanding: mammals in burrows, birds, crocodilians, turtles, some sharks, insects, the detritivores that could live on dead matter while the Sun stayed dark.

Recovery was slow and uneven. Depending on the group, ecosystems took anywhere from two to ten million years to rebuild, and mammals diversified rapidly, in geological terms, in the aftermath.

More finely, marine ecosystems needed about two to four million years to fully recover, and the first few hundred thousand years belonged to opportunistic disaster taxa, fern spikes on land and small generalists in the sea. Mammalian diversification itself did not accelerate significantly until two to three hundred thousand years after the impact; the emptied world took time to fill.

The Deccan Traps in western India are a flood-basalt province covering something like five hundred thousand square kilometers. The eruptions began roughly a quarter of a million years before Chicxulub and continued for about half a million years after it, releasing enormous volumes of carbon dioxide and sulfur dioxide. That is more than enough to stress the global climate on its own, with warming and acid rain, before the asteroid ever arrived.

This gave rise to the press-pulse model, proposed by Arens and West in 2008: the Deccan volcanism was the long, grinding press that weakened ecosystems, and Chicxulub was the sudden pulse that tipped them over. A minority of researchers, most prominently Gerta Keller, argue the balance the other way, making Deccan volcanism the primary cause and the impact a secondary blow. The research documents label this a minority position, and it is worth understanding why.

Keller's stronger version rests on a reading of the El Penon section in Mexico that places the impact about three hundred thousand years before the extinction. High-precision radiometric dating has not supported that gap. Work by Renne and colleagues in 2013, and since, has established that the impact and the extinction boundary are synchronous, and Hull and colleagues showed in 2020 that marine ecosystems show no major turnover during even the most intense pre-impact Deccan eruptions. The biological collapse begins at the impact horizon, not before it.

There is also a twist that ties the two suspects together. Studies in 2015 and 2019 found that the Deccan eruptions actually accelerated within tens of thousands of years of the impact, as if the seismic shock had shaken the volcanic plumbing into overdrive, a literal one-two punch. Antipodal focusing, the idea that impact energy converged on the far side of the planet, is one proposed mechanism.

The honest bottom line, and the one the research documents settle on, is not a coin toss. Both the impact and the Deccan Traps contributed to the environmental stress of the end-Cretaceous world. But the impact was the primary trigger of the mass extinction. The volcano loaded the gun; the asteroid pulled the trigger.

With the dinosaurs gone, the mammals that had spent a hundred million years small and nocturnal inherited an emptied world and radiated into it. The blunt version, which the research supports, is that without the K-Pg extinction, mammals, and therefore humans, would very likely never have risen to ecological dominance at all. We are, in a real sense, the beneficiaries of a catastrophe.

Stephen Jay Gould pushed this further into a famous argument about contingency: rerun the tape of life, let the asteroid miss, and the dinosaurs keep the planet while mammals stay small forever, so nothing like us ever appears. It is a genuinely speculative claim, unprovable in either direction. The counter-case is real too, that mammals were already diversifying in the late Cretaceous and might have pressed the dinosaurs eventually regardless. The question is unfalsifiable, and it is also one of the most quietly staggering thoughts in all of paleontology.

Some models of the crater's shape hint that the impactor may have been a binary asteroid, two bodies orbiting each other rather than one. It is a real reading of the morphology, and it is unproven.

There is also a second, much smaller crater in the mix. The Boltysh crater in Ukraine, about 24 kilometers across, formed close enough to the K-Pg boundary that it has been proposed as a companion, or even as a precursor that stressed ecosystems before the main blow. But how close is exactly the problem. Our own research documents give its timing relative to Chicxulub three different ways: one puts the gap at just two to five thousand years, another at around eight hundred thousand years before, a third at several hundred thousand. Those figures differ by more than two orders of magnitude, so any claim that Boltysh and Chicxulub were a linked pair has to be treated as unresolved, not established.

A larger claim, the so-called Shiva crater, a proposed five-hundred-kilometer structure off the west coast of India, has been put forward by Sankar Chatterjee as a second, even bigger K-Pg impact. Most geophysicists read that structure as volcanic or tectonic in origin, and the geological consensus treats it as debunked rather than confirmed.

Further out still is the idea that impacts come on a schedule. Randall and Reece proposed in 2014 that the Solar System's passage through a disk of dark matter in the galactic plane could periodically fling Oort Cloud comets inward, producing a roughly thirty-million-year rhythm of impacts. It is a genuinely creative hypothesis, and highly speculative, resting on a periodicity in the extinction record that is itself disputed. Related proposals invoke a hidden companion star, Nemesis, or a distant Planet X. The impact record may simply be too patchy to show a cycle at all.

On the other side of the ledger sit the claims the evidence rules out. The idea that the K-Pg extinction was caused purely by volcanism, with the Yucatan structure not an impact crater at all, is flatly contradicted by the physical record: the shocked quartz, the global iridium, the crater's own morphology. Officer and Drake mounted a serious challenge to the impact hypothesis back in 1985, but their objections have been answered point by point by the decades of work since.

The stronger version of the volcanic argument, associated with Gerta Keller, that the impact was only a minor factor and the extinction was primarily volcanic, runs straight into the sharp, geologically instantaneous extinction signal recorded at the boundary worldwide. As a call for the impact's importance to be weighed honestly against the Deccan Traps, the Keller camp has kept the field careful. As a claim that the asteroid was a bit player, it is not supported by the weight of the evidence.

And no, the dinosaurs did not quietly survive. Claims of dinosaur fossils from above the boundary surface from time to time, but none has ever survived peer review; every one is explained by reworking, older bones eroded out and redeposited in younger sediment. The one honest caveat is the one that sounds like a technicality but is not: birds are dinosaurs, the living branch of the family. Every non-avian lineage, though, ended at the boundary and stayed ended.