Mass Extinctions and Their Consequences
Sixty-six million years ago, a rock roughly 10 km wide struck the shallow sea off what is now the Yucatán Peninsula. Within hours, the impact triggered fires, tsunamis, and a global blackout of sunlight. Within two million years, roughly 75% of species on Earth were gone. This included every non-avian dinosaur. Yet from that devastation came a diversification. It gave rise to nearly every mammal lineage alive today, including our own. Mass extinctions are the great punctuation marks of life's history. They are brief geological moments when extinction rates spike far above the background rate. In these moments, a large fraction of species across many unrelated groups vanish simultaneously. They matter to evolutionary biology because they show that the rules of survival can change abruptly. During ordinary times, natural selection favors traits that improve competitive ability, reproduction, or predator avoidance within a stable ecological context. During a mass extinction, an entirely different filter operates. This filter often has little to do with how "fit" a species was under normal conditions. Understanding this distinction, selectivity by extinction regime, is essential to understanding why life's history looks the way it does. This article builds on the general fossil record and the tempo of macroevolutionary change. It focuses on the five most severe extinction events preserved in the rock record, the "Big Five," plus the ongoing sixth. It explains the physical and biological mechanisms that cause mass die-offs. It also explains the evidence that pins each event to a specific cause. It covers how selectivity differs from background extinction. Finally, it shows how survivors reorganize into new adaptive radiations once ecological space empties out. The story connects directly to later topics. The empty niches left by extinction are a primary engine of Convergent Evolution and Homoplasy. One mass extinction in particular set the stage for the primate and hominin lineage discussed in Human Evolution and the Hominin Lineage. The deep lesson is that extinction is not simply the opposite of adaptation. It is a distinct evolutionary process with its own rules. It has repeatedly reset the direction of life on Earth.
- Mass extinctions are brief intervals where global extinction rates rise far above background levels, eliminating a large fraction of species across unrelated taxonomic groups.
- The "Big Five" are the end-Ordovician, Late Devonian, end-Permian, end-Triassic, and end-Cretaceous events; a sixth, human-driven extinction is underway now.
- Causes include glaciation and sea-level change, massive volcanism (flood basalts), asteroid impact, and ocean anoxia — often acting in combination.
- Extinction selectivity during mass events often differs from selectivity during normal (background) times, favoring traits like geographic range and physiological tolerance over local competitive superiority.
- Survivors radiate into vacated ecological space, producing adaptive radiations; the end-Cretaceous extinction directly enabled the diversification of placental mammals.
- Evidence for causes comes from convergent geochemical, sedimentological, and paleontological signatures — iridium anomalies, carbon isotope excursions, and precise radiometric dating of extinction horizons.
Mass extinctions are rare, rapid, and non-random pulses of species loss that empty ecological space and thereby redirect the subsequent course of evolution more powerfully than millions of years of ordinary natural selection.
The core idea
An extinction is the permanent disappearance of a species; the background extinction rate is the low, roughly steady rate at which species disappear during normal geological intervals, estimated from the fossil record at somewhere between 0.1 and 1 extinctions per million species-years for well-studied marine invertebrates. A mass extinction is instead defined operationally: a geologically brief interval (typically under 2 million years, sometimes far less) during which a large proportion of taxa — conventionally taken as at least 50–75% of species — disappear across geographically widespread and taxonomically unrelated groups. The key diagnostic is not just the number of extinctions but their breadth (many unrelated clades affected simultaneously), rate (compressed into a short window relative to typical turnover), and global rather than regional extent. Paleontologists led by Jack Sepkoski and David Raup, compiling a database of marine animal families and genera through the 1970s–1980s, identified five intervals in the last 540 million years where extinction intensity spiked far above the background trend: these are the Big Five — end-Ordovician (~445 Ma), Late Devonian (~372 Ma), end-Permian (~252 Ma), end-Triassic (~201 Ma), and end-Cretaceous (~66 Ma). A crucial conceptual distinction, developed especially by David Jablonski, is between extinction selectivity in background versus mass-extinction regimes. During background times, traits that matter for local, ecological competition (efficient resource use, predator avoidance, rapid reproduction) tend to determine which species persist. During mass extinctions, different traits often matter more: broad geographic range, eurytopy (tolerance of many environments), and physiological robustness to the specific stress (heat, anoxia, acidification) driving the event. This means a species' fitness under normal conditions is a poor predictor of its survival during a crisis — a central reason mass extinctions redirect the tree of life rather than simply pruning its weakest branches.
Mechanism
Step 1 — A perturbation exceeds the resilience of ecosystems. Every extinction event begins with an environmental driver operating at a magnitude or speed that overwhelms species' normal buffering capacity: physiological tolerance, migration, or local adaptation. The known drivers recur across the Big Five in different combinations:
| Driver | Mechanism of kill | Example event |
|---|---|---|
| Flood-basalt volcanism | CO₂/SO₂ release → warming, acid rain, ocean acidification, anoxia | end-Permian (Siberian Traps), end-Triassic (CAMP) |
| Asteroid/comet impact | Dust/soot blocks sunlight → collapse of photosynthesis → food-web collapse | end-Cretaceous (Chicxulub) |
| Glaciation & sea-level fall | Habitat loss on continental shelves, cooling | end-Ordovician |
| Ocean anoxia | Loss of dissolved O₂ in deep and shelf waters, kills benthic and pelagic life | Late Devonian, end-Permian, end-Triassic (contributing) |
| Ocean acidification | Impairs calcification in shelled/skeletal organisms | end-Permian, end-Cretaceous |
Step 2 — The perturbation cascades through food webs and biogeochemical cycles. A single trigger rarely kills directly; instead it destabilizes the base of ecosystems. For instance, sunlight-blocking dust after an impact suppresses photosynthesis on land and in the surface ocean; primary producers crash; herbivores and phytoplankton grazers starve; predators at the top of the food chain follow. Large-bodied, energetically demanding animals requiring large, stable food supplies (like non-avian dinosaurs) are disproportionately vulnerable to this kind of cascading famine, while small-bodied, detritivorous, or omnivorous animals that can survive on scarce or stored resources persist better.
Step 3 — Selectivity operates on standing variation, not on new adaptation. Because mass extinctions unfold over thousands to a few hundred thousand years — far too fast for new adaptations to evolve in response — survival is determined by which pre-existing traits happen to confer resilience. Traits found repeatedly to correlate with survival include:
- Wide geographic range (a species is less likely to be extirpated everywhere at once)
- Broad environmental tolerance (eurytopy) — generalists over specialists
- Small body size and low metabolic demand
- Ability to enter dormancy, burrow, or use refugia (freshwater, deep soil, polar regions)
- Non-reliance on live photosynthetic material (detritivores and scavengers over herbivores in "photosynthetic shutdown" events)
Step 4 — Extinction empties ecological space (niche vacancy). Once dominant clades are removed, their ecological roles (large herbivore, apex predator, reef-builder) become vacant, even though the physical resources and habitats persist.
Step 5 — Surviving lineages diversify into vacated niches — adaptive radiation. With competitors and predators removed, surviving lineages undergo rapid speciation and morphological diversification as they evolve to exploit newly available ecological space. This is the direct evolutionary consequence of mass extinction: it does not merely subtract diversity, it resets the starting conditions for subsequent diversification, often determining which major body plans and lineages come to dominate for tens of millions of years afterward.
The evidence
Establishing the cause of a mass extinction requires converging lines of geochemical, sedimentological, and paleontological evidence, precisely dated and cross-correlated globally. The best-documented case is the Cretaceous–Paleogene (K–Pg) extinction, 66 million years ago.
The Alvarez hypothesis and the iridium anomaly (1980). Physicist Luis Alvarez, his son geologist Walter Alvarez, and chemists Frank Asaro and Helen Michel examined a thin clay layer marking the Cretaceous–Paleogene boundary at Gubbio, Italy. They measured iridium, an element rare in Earth's crust but relatively abundant in asteroids and comets, and found concentrations at the boundary layer roughly 30 times background levels; equivalent enrichment was later found at boundary sections worldwide, including Stevns Klint in Denmark and sites in New Zealand and the Pacific. Because iridium is so rare terrestrially, its global, synchronous spike strongly implied an extraterrestrial impact rather than a local volcanic or sedimentary artifact. The Alvarez team calculated that an impactor roughly 10 km in diameter would deliver enough iridium and enough energy to account for both the geochemical signal and a global extinction event.
Finding the crater. The hypothesis required an impact site of the right age and size. In the late 1970s and 1980s, geophysicists working for Mexican oil exploration (notably Glen Penfield and Antonio Camargo) had already identified a buried circular gravity and magnetic anomaly beneath the Yucatán Peninsula near the town of Chicxulub, roughly 180 km across — but its significance was not connected to the extinction until the early 1990s, when drill cores recovered shocked quartz and impact-melt rock dated to precisely 66 million years old, matching the boundary layer.
Shocked quartz and tektites. Independent confirmation came from shocked quartz grains (quartz crystals bearing microscopic deformation lamellae that form only under the extreme, near-instantaneous pressures of impact or nuclear detonation, not from volcanism) found in the boundary clay across North America, and from tektites (small glass spherules formed from molten rock ejected by the impact and cooled in flight), found embedded in boundary sediments in Haiti and the Gulf of Mexico region.
High-precision dating. Radiometric dating using argon-argon and uranium-lead methods on impact glass and volcanic ash beds bracketing the boundary has pinned the Chicxulub impact and the mass extinction to the same date within a resolution of a few tens of thousands of years — a 2013 study by Paul Renne and colleagues dated the impact to 66.043 ± 0.043 million years ago, statistically indistinguishable from the extinction horizon, supporting a causal link rather than coincidence.
The debate over volcanism. Not all researchers attribute the K–Pg extinction solely to impact. The Deccan Traps, a vast flood-basalt province in India, were erupting large volumes of lava and releasing climate-altering gases in the several hundred thousand years surrounding the boundary. Current consensus, supported by high-resolution dating (work by Blair Schoene and colleagues), holds that Deccan volcanism was already stressing global ecosystems and climate before the impact, but that the Chicxulub impact was the abrupt trigger of the mass extinction itself, with post-impact volcanism possibly slowing biotic recovery. This is a genuinely contested point of emphasis, and it illustrates that extinction causation is often multifactorial rather than singular.
The end-Permian extinction — the largest of the Big Five, eliminating an estimated 81% of marine species and about 70% of terrestrial vertebrate genera — is dated with high precision (U-Pb zircon dating) to have occurred in less than 200,000 years, coincident with the eruption of the Siberian Traps, a flood basalt province covering an area comparable to modern Europe. Evidence includes a sharp, globally correlated negative excursion in carbon-13 isotopes (indicating massive release of light carbon from volcanic CO₂ and possibly destabilized methane clathrates), fossil and geochemical evidence of widespread ocean anoxia (pyrite framboids, biomarkers for anoxic sulfur bacteria), and evidence of ocean acidification recorded in the calcification patterns of surviving marine organisms.
Worked examples
Example 1 — Reading selectivity from body-size data across the K–Pg boundary. Suppose a paleontologist surveys terrestrial vertebrate genera straddling the K–Pg boundary in a well-sampled formation and finds:
| Body mass class | Genera present before boundary | Genera surviving after |
|---|---|---|
| >25 kg | 24 | 1 |
| 1–25 kg | 40 | 9 |
| <1 kg | 30 | 18 |
Survival proportions: >25 kg = 1/24 ≈ 4%; 1–25 kg = 9/40 = 22.5%; <1 kg = 18/30 = 60%. This monotonic decline in survival with increasing body size is exactly the pattern documented across the K–Pg boundary and is mechanistically consistent with the "impact winter" model: large animals require large, stable food supplies and long generation times, making them vulnerable to a multi-year collapse in primary productivity, while small animals with lower absolute energy requirements, faster reproduction, and access to detritus, seeds, or insects (food sources that persist even when photosynthesis is suppressed) are disproportionately likely to survive. This is a textbook illustration of extinction-regime selectivity: body size, largely irrelevant as a direct target of selection under background conditions, becomes strongly predictive of survival during a mass extinction.
Example 2 — Estimating extinction magnitude and its effect on diversification rate. Suppose a marine clade has 500 genera immediately before an extinction event, and the fossil record shows only 60 genera crossing into the recovery interval afterward. Percent extinction = (500 − 60)/500 × 100 = 88% genus-level loss. Now suppose that in the 10 million years following the extinction, genus richness in this clade rebounds to 400. We can estimate an average net diversification rate r using the exponential growth model: N(t) = N₀ · e^(rt) 400 = 60 · e^(r × 10) 6.667 = e^(10r) ln(6.667) = 10r r ≈ 1.897 / 10 ≈ 0.19 per million years Compare this to typical background net diversification rates for marine invertebrate genera, often estimated around 0.05–0.10 per million years. The roughly two- to fourfold elevation in diversification rate reflects the classic post-extinction pattern: with competitors removed and niche space empty, surviving lineages diversify unusually fast — exactly the dynamic documented for surviving bivalve and gastropod clades after the end-Permian extinction, and for placental mammals after the K–Pg extinction, where molecular and fossil evidence together show a marked acceleration in lineage splitting during the first 10 million years of the Paleocene.
Common misconceptions & traps
| Misconception | Correction |
|---|---|
| "Mass extinctions kill the least fit species." | Survival is often determined by traits irrelevant to normal fitness (range size, body size, tolerance), not by competitive superiority under background conditions. |
| "Dinosaurs were already declining and doomed regardless of the asteroid." | Evidence for a long-term Cretaceous decline in dinosaur diversity is debated and largely not supported by recent high-resolution fossil analyses; the abruptness of the K–Pg boundary extinction strongly implicates the impact as the proximate cause. |
| "Only one cause (e.g., the asteroid) explains each extinction." | Most Big Five events involve multiple interacting stressors (e.g., Deccan volcanism plus impact at K–Pg; Siberian Traps plus anoxia and acidification at end-Permian). |
| "Mass extinctions are bad for evolution." | Extinction removes diversity but also releases ecological opportunity, often accelerating subsequent diversification and enabling major evolutionary innovations. |
| "Evolution 'directs' surviving lineages toward improvement." | Survivors are not "better," only differently suited to the specific stress; radiation afterward is undirected, opportunistic diversification into vacant niches, not progress toward a goal. |
The scientist behind it
Walter Alvarez was a field geologist working on the stratigraphy of Italian limestone sequences in the 1970s when he became puzzled by a thin, distinctive clay layer separating Cretaceous from Paleogene limestone beds at Gubbio — a boundary marking the sudden disappearance of numerous foraminifera species in the fossil record just below it. Seeking a way to measure how long this layer took to accumulate, he consulted his father, Nobel laureate physicist Luis Alvarez, who suggested measuring the rare element iridium, delivered to Earth at a slow, roughly constant rate by cosmic dust, as a "clock." The unexpectedly enormous iridium spike they found, far too large to be explained by slow deposition, redirected the entire investigation: rather than measuring time, they had discovered evidence of a catastrophic event. Their 1980 paper, co-authored with chemists Frank Asaro and Helen Michel, proposed the asteroid impact hypothesis — initially met with skepticism from many paleontologists and volcanologists, and vindicated over the following decade as shocked quartz, tektites, and eventually the Chicxulub crater itself were identified.
Origins · 1 minWalter Alvarez and how this was discovered
Where this matters today
Mass extinction research is not an antiquarian pursuit — it directly informs how biologists understand and respond to the modern biodiversity crisis, often called the "sixth extinction." Current species loss rates, driven by habitat destruction, climate change, invasive species, and overexploitation, are estimated by many biologists to exceed background extinction rates by one to several orders of magnitude, raising the question of whether human impacts constitute a true mass extinction in the geological sense. Understanding extinction selectivity helps conservation biologists identify which traits — small geographic range, narrow habitat specialization, low reproductive rate, large body size — predict elevated extinction risk today, directly shaping IUCN Red List risk assessments and the prioritization of conservation resources. In paleoclimatology and public policy, the end-Permian extinction serves as the closest deep-time analog for anthropogenic ocean acidification and warming, since it involved a massive, rapid release of volcanic CO₂; researchers use its recorded sequence of anoxia, acidification, and marine die-off, reconstructed from isotope geochemistry, to calibrate models projecting the ecological consequences of modern fossil-fuel CO₂ release. In agriculture and biotechnology, understanding how surviving lineages tolerated past environmental extremes (heat, drought, altered atmospheric chemistry) informs the search for stress-tolerance genes in crop wild relatives. In medicine and epidemiology, the general principle that catastrophic, rapid-onset selective regimes favor different traits than gradual ones is applied in modeling pathogen and host population crashes and bottleneck events. Finally, in evolutionary developmental biology and biogeography, mass extinctions remain the standard explanatory framework for why certain major clades (placental mammals, teleost fishes, angiosperms) came to dominate the modern biosphere — not because they were inherently superior, but because they were positioned, by chance and physiology, to survive a crisis and then radiate into the ecological vacuum it left behind.
Ask the tutor: How do scientists distinguish whether an extinction was caused by volcanism, impact, or both, when both leave overlapping geochemical signatures at the same boundary?
Were all five of the "Big Five" extinctions caused by asteroid impacts?
No. Only the end-Cretaceous (K–Pg) extinction has strong, well-corroborated evidence for a large asteroid impact as the primary trigger, via the iridium anomaly, shocked quartz, and the dated Chicxulub crater. The end-Permian and end-Triassic extinctions are most strongly linked to massive flood-basalt volcanism (the Siberian Traps and the Central Atlantic Magmatic Province, respectively), which released enormous volumes of CO₂ and SO₂, driving warming, acidification, and anoxia. The end-Ordovician extinction is linked to glaciation and sea-level fall, and the Late Devonian event is associated with ocean anoxia, possibly linked to the spread of early forests altering nutrient runoff. Causes differ across events, though volcanism and anoxia recur as common threads.
Why did small-bodied animals survive the K–Pg extinction better than large ones?
The leading explanation is the "impact winter" hypothesis: dust, soot, and sulfate aerosols from the Chicxulub impact blocked sunlight for months to years, collapsing photosynthesis on land and in surface oceans. Large-bodied herbivores and predators, which require large, stable, continuously replenished food supplies and have slow reproductive rates, could not survive a prolonged collapse in primary productivity. Small-bodied animals, with lower absolute energy needs, faster reproduction, and the ability to subsist on dormant seeds, detritus, insects, or scavenged material (resources less dependent on ongoing photosynthesis), were disproportionately likely to survive and later diversify.
How do scientists know the timing of the Chicxulub impact matches the extinction so precisely?
Researchers use radiometric dating methods, particularly argon-argon dating on impact glass (tektites and impact melt) and uranium-lead dating on zircon crystals in volcanic ash layers that bracket the boundary in multiple locations worldwide. A landmark 2013 study by Paul Renne and colleagues dated the Chicxulub impact melt and the K–Pg boundary ash layer to within about 32,000 years of each other, essentially simultaneous on geological timescales, and consistent with the abrupt disappearance of many fossil taxa recorded in the same rock layers. This convergence of independent dating methods across multiple sites is what elevates the impact-extinction link from correlation to strong causal inference.
What is meant by "extinction selectivity," and why does it matter for evolutionary theory?
Extinction selectivity refers to the non-random pattern of which species or lineages survive a given extinction episode, based on shared traits. David Jablonski's work showed that the traits favored during mass extinctions (wide geographic range, environmental tolerance, small body size) often differ from the traits favored by ordinary natural selection during stable times (competitive ability, specialization, efficient resource use). This matters because it means mass extinctions are not simply an intensified version of everyday selection; they represent a distinct evolutionary regime that can override the outcomes of long-term background adaptation, explaining why some highly successful, well-adapted lineages (like non-avian dinosaurs) vanish while less dominant lineages persist and later flourish.
Did the extinction of the dinosaurs directly "cause" the rise of mammals?
Not in a directed sense, but the extinction was a necessary precondition. Mammals had already existed for over 150 million years before the K–Pg extinction, mostly as small-bodied, nocturnal or burrowing animals, likely constrained ecologically by dinosaur dominance of large-bodied niches. Once non-avian dinosaurs and many other dominant reptile lineages were eliminated, the ecological space for large-bodied terrestrial herbivores and predators became vacant. Molecular and fossil evidence together show a rapid diversification of placental mammal lineages in the following few million years, filling those vacated niches. The extinction did not "aim" to produce mammals; it removed constraints, and mammalian diversification followed as opportunistic, undirected adaptive radiation.
Is the current biodiversity crisis really comparable to the Big Five?
It is debated, and the comparison must be made carefully. Current species extinction rates, based on documented extinctions and extrapolated habitat loss, are estimated by many researchers to be tens to hundreds of times higher than typical background rates, and possibly approaching rates seen in past mass extinctions if current trends continue for centuries. However, the Big Five each eliminated 50–90% of species over intervals from centuries to under two million years, a scale not yet reached today. Most conservation biologists treat the comparison as a serious warning rather than a settled equivalence, using deep-time extinction selectivity patterns (range size, specialization, body size) to help predict which modern species and ecosystems are at greatest risk.
Why is body size such a recurring theme in extinction selectivity across different events?
Large body size correlates with several traits that become liabilities during rapid environmental crises: higher absolute energy requirements, longer generation times, smaller population sizes, and often narrower dietary or habitat specialization. When food webs collapse rapidly, as after the K–Pg impact or during severe anoxic events, these traits reduce the buffering capacity that helps a population survive short-term resource scarcity. This pattern has been documented independently across multiple, mechanistically distinct extinction events (impact winter, volcanic-driven anoxia, glaciation), suggesting it reflects a general physiological constraint on resilience rather than a coincidence specific to any one crisis.