An extinction event (also known as: mass extinction; extinction-level event, ELE) is a sharp decrease in the number of species in a relatively short period of time. Mass extinctions affect most major taxonomic groups present at the time — birds, mammals, reptiles, amphibians, fish, invertebrates and other simpler life forms. They may be caused by one or both of:
Over 99% of species that ever lived are now extinct, but extinction occurs at an uneven rate. Based on the fossil record, the background rate of extinctions on Earth is about two to five taxonomic families of marine invertebrates and vertebrates every million years. Marine fossils are mostly used to measure extinction rates because they are more plentiful and cover a longer time span than fossils of land organisms.
Since life began on earth, several major mass extinctions have significantly exceeded the background extinction rate. The most recent, the Cretaceous–Tertiary extinction event, occurred 65 million years ago, and has attracted more attention than all others because it killed the dinosaurs. In the past 540 million years there have been five major events when over 50% of animal species died. There probably were mass extinctions in the Archean and Proterozoic Eons, but before the Phanerozoic there were no animals with hard body parts to leave a significant fossil record.
Estimates of the number of major mass extinctions in the last 540 million years range from as few as five to more than twenty. These differences stem from the threshold chosen for describing an extinction event as "major", and the data chosen to measure past diversity.
The classical "Big Five" mass extinctions identified by Jack Sepkoski and David M. Raup in their 1982 paper are widely agreed upon as some of the most significant: End Ordovician, Late Devonian, End Permian, End Triassic, and End Cretaceous.  In addition, there is mounting evidence for two other mass extinctions of a similar scale: the current Holocene extinction event and the end-Ediacaran extinction at the start of the Phanerozoic eon.
These and a selection of other extinction events are outlined below. The articles about individual mass extinctions describe their effects in more detail and discuss theories about their causes.
The older the fossil record gets the more difficult it is to read it. This is because:
It has been suggested that the apparent variations in marine biodiversity may actually be an artifact, with abundance estimates directly related to quantity of rock available for sampling from different time periods. However, statistical analysis shows that this can only account for 50% of the observed pattern, and other evidence (such as fungal spikes) provides reassurance that most widely accepted extinction events are indeed real. A quantification of the rock exposure of Western Europe does indicate that many of the minor events for which a biological explanation has been sought are most readily explained by sampling bias.
Minor extinction events include:
Quaternary Period (disputed)
Mass extinctions have sometimes accelerated the evolution of life on earth. When dominance of particular ecological niches passes from one group of organisms to another, it is rarely because the new dominant group is "superior" to the old and usually because an extinction event eliminates the old dominant group and makes way for the new one. 
For example mammaliformes ("almost mammals") and then mammals existed throughout the reign of the dinosaurs, but could not compete for the large terrestrial vertebrate niches which dinosaurs monopolized. The end-Cretaceous mass extinction removed the non-avian dinosaurs and made it possible for mammals to expand into the large terrestrial vertebrate niches.
Another point of view put forward in the Escalation hypothesis predicts that species in ecological niches with more organism to organism conflict will be less likely to survive extinctions. This is because the very traits that keep a species numerous and viable under fairly static conditions become a burden once population levels fall among competing organisms during the dynamics of an extinction event.
Furthermore, many groups which survive mass extinctions do not recover in numbers or diversity, and many of these go into long-term decline, and these are often referred to as "Dead Clades Walking". So analysing extinctions in terms of "what died and what survived" often fails to tell the full story.
The gaps between mass extinctions appear to be becoming longer, while the average and background rates of extinction are decreasing. Mass extinctions are thought to result when a long-term stress is compounded by a short term shock. Over the course of the Phanerozoic, individual taxa appear to be less likely to become extinct at any time, which may reflect more robust food webs as well as less extinction-prone species and other factors such as continental distribution. However the taxonomic susceptibility to extinction does not appear to make mass extinctions more or less probable.
The idea that mass extinctions are becoming less frequent is rather speculative – from a statistical point of view a sample of about 10 extinction events is too small to be a reliable sign of any actual trend. But the average and background rates of extinction are based on hundreds of samples over a period of 550M years, so the apparent decrease in these rates is statistically significant and needs to be explained.
Both of these phenomena could be explained in one or more ways:
There is still debate about the causes of all mass extinctions before the Holocene. In general, large extinctions result when a biosphere under long term stress (such as climate change) undergoes a short term shock (such as a meteor impact).
A good theory for a particular mass extinction should: (i) explain all of the losses, not just focus on a few groups (such as dinosaurs); (ii) explain why particular groups of organisms died out and why others survived; (iii) provide mechanisms which are strong enough to cause a mass extinction but not a total extinction; (iv) be based on events or processes that can be shown to have happened, not just inferred from the extinction.
It may be necessary to consider combinations of causes. For example the marine aspect of the end-Cretaceous extinction appears to have been caused by several processes which partially overlapped in time and may have had different levels of significance in different parts of the world.
Arens and West (2006) proposed a "press / pulse" model in which mass extinctions generally require two types of cause: long-term pressure on the eco-system ("press") and a sudden catastrophe ("pulse") towards the end of the period of pressure. Their statistical analysis of marine extinction rates throughout the Phanerozoic suggested that neither long-term pressure alone nor a catastrophe alone was sufficient to cause a significant increase in the extinction rate.
Macleod (2001) summarised the relationship between mass extinctions and events which are most often cited as causes of mass extinctions, using data from Courtillot et al (1996), Hallam (1992) and Grieve et al (1996):
The most commonly suggested causes of mass extinctions are listed below
The formation of large igneous provinces by flood basalt events could have:
Flood basalt events occur as pulses of activity punctuated by dormant periods. As a result they are likely to cause the climate to oscillate between cooling and warming, but with an overall trend towards warming as the carbon dioxide they emit can stay in the atmosphere for hundreds of years. Massive volcanism caused or contributed to the End-Cretaceous, End-Permian, End Triassic and End Jurassic extinctions.
These are often clearly marked by world-wide sequences of contemporaneous sediments which show all or part of a transition from sea-bed to tidal zone to beach to dry land - and where there is no evidence that the rocks in the relevant areas were raised by geological processes such as orogeny. Sea-level falls could reduce the continental shelf area (the most productive part of the oceans) sufficiently to cause a marine mass extinction, and could disrupt weather patterns enough to cause extinctions on land. But sea-level falls are very probably the result of other events, such as sustained global cooling or the sinking of the mid-ocean ridges.
A study, published in the journal Nature (online June 15, 2008) established a relationship between the speed of mass extinction events and changes in sea level and sediment. The study suggests changes in ocean environments related to sea level exert a driving influence on rates of extinction, and generally determine the composition of life in the oceans.
The impact of a sufficiently large asteroid or comet could have caused food chains to collapse both on land and at sea by producing dust and particulate aerosols and thus inhibiting photosynthesis. Impacts on sulfur-rich rocks could have emitted sulfur oxides precipitating as poisonous acid rain, contributing further to the collapse of food chains. Such impacts could also have caused megatsunamis and / or global forest fires, but evidence for these events has been difficult to prove.
Only the Cretaceous–Tertiary extinction event is associated with strong evidence of such an impact, but that impact is easily the largest for which there is strong evidence.
Sustained global cooling could kill many polar and temperate species and force others to migrate towards the equator; reduce the area available for tropical species; often make the Earth's climate more arid on average, mainly by locking up more of the planet's water in ice and snow. The glaciation cycles of the current ice age are believed to have had only a very mild impact on biodiversity, so the mere existence of a significant cooling is not sufficient on its own to explain a mass extinction.
It has been suggested that global cooling caused or contributed to the End-Ordovician, Permian-Triassic, Late Devonian extinctions, and possibly others. Sustained global cooling is distinguished from the temporary climatic effects of flood basalt events or impacts.
This would have the opposite effects: expand the area available for tropical species; kill temperate species or force them to migrate towards the poles (or perish); possibly cause severe extinctions of polar species; often make the Earth's climate wetter on average, mainly by melting ice and snow and thus increasing the volume of the water cycle. It might also cause anoxic events in the oceans (see below).
Global warming as a cause of mass extinction is supported by several recent studies.
The most dramatic example of sustained warming is the Paleocene-Eocene Thermal Maximum, which was associated with one of the smaller mass extinctions. It has also been suggested to have caused the Triassic-Jurassic extinction event, during which 20% of all marine families went extinct. Furthermore, the Permian–Triassic extinction event has been suggested to have been caused by warming.   
See main article: Clathrate Gun Hypothesis.
Clathrates are composites in which a lattice of one substance forms a cage round another. Methane clathrates (in which water molecules are the cage) form on continental shelves. These clathrates are likely to break up rapidly and release the methane if the temperature rises quickly or the pressure on them drops quickly — for example in response to sudden global warming or a sudden drop in sea level or even earthquakes. Methane is a much more powerful greenhouse gas than carbon dioxide, so a methane eruption ("clathrate gun") could cause rapid global warming or make it much more severe if the eruption was itself caused by global warming.
The most likely signature of such a methane eruption would be a sudden decrease in the ratio of carbon-13 to carbon-12 in sediments, since methane clathrates are low in carbon-13; but the change would have to be very large, as other events can also reduce the percentage of carbon-13.
It has been suggested that "clathrate gun" methane eruptions were involved in the end-Permian extinction ("the Great Dying") and in the Paleocene-Eocene Thermal Maximum, which was associated with one of the smaller mass extinctions.
Anoxic events are situations in which the upper and even the middle layers of the ocean become deficient or totally lacking in oxygen. Their causes are complex and controversial, but all known instances are associated with severe and sustained global warming, mostly caused by massive sustained volcanism.
It has been suggested that anoxic events caused or contributed to the late Devonian, Permian-Triassic and Triassic-Jurassic extinctions. On the other hand, there are widespread black shale beds from the mid-Cretaceous which indicate anoxic events but are not associated with mass extinctions.
Kump, Pavlov and Arthur (2005) have proposed that during the Permian-Triassic extinction event the warming also upset the oceanic balance between photosynthesising plankton and deep-water sulfate-reducing bacteria, causing massive emissions of hydrogen sulfide which poisoned life on both land and sea and severely weakened the ozone layer, exposing much of the life that still remained to fatal levels of UV radiation.  
Oceanic overturn is a disruption of thermo-haline circulation which lets surface water (which is more saline than deep water because of evaporation) sink straight down, bringing anoxic deep water to the surface and therefore killing most of the oxygen-breathing organisms which inhabit the surface and middle depths. It may occur either at the beginning or the end of a glaciation, although an overturn at the start of a glaciation is more dangerous because the preceding warm period will have created a larger volume of anoxic water.
Unlike other oceanic catastrophes such as regressions (sea-level falls) and anoxic events, overturns do not leave easily-identified "signatures" in rocks and are theoretical consequences of researchers' conclusions about other climatic and marine events.
A nearby gamma ray burst (fewer than 6000 light years away) could sufficiently irradiate the surface of Earth to kill organisms living there and destroy the ozone layer in the process. From statistical arguments, approximately 1 gamma ray burst would be expected to occur close to Earth in the last 540 million years. A proposal that a supernova or gamma ray burst had caused a mass extinction would also have to be backed up by astronomical evidence of such an explosion at the right place and time.
It has been suggested that a supernova or gamma ray burst caused the End-Ordovician extinction.
Movement of the continents into some configurations can cause or contribute to extinctions in several ways: by initiating or ending ice ages; by changing ocean and wind currents and thus altering climate; by opening seaways or land bridges which expose previously isolated species to competition for which they are poorly-adapted (for example the extinction of most American marsupials after the creation of a land bridge between North and South America). Occasionally continental drift creates a super-continent which includes the vast majority of Earth's land area, which in addition to the effects listed above is likely to reduce the total area of continental shelf (the most species-rich part of the ocean) and produce a vast, arid continental interior which may have extreme seasonal variations.
It is widely thought that the creation of the super-continent Pangaea contributed to the End-Permian mass extinction. Pangaea was almost fully formed at the transition from mid-Permian to late-Permian, and the "Marine genus diversity" diagram at the top of this article shows a level of extinction starting at that time which might have qualified for inclusion in the "Big Five" if it were not overshadowed by the "Great Dying" at the end of the Permian.
Plate tectonics is the mechanism which drives many of the possible causes of mass extinctions, especially volcanism and continental drift. So it is implicated in many extinctions, but in each case it is necessary to specify which manifestations of plate tectonics were involved.
Many other hypotheses have been proposed, such as the spread of a new disease or simple out-competition following an especially successful biological innovation. But all have been rejected, usually for one of the following reasons: they require events or processes for which there is no evidence; they assume mechanisms which are contrary to the available evidence; they are based on other theories which have been rejected or superseded.
It has been suggested by several sources that biodiversity and/or extinction events may be influenced by cyclic processes. The best-known hypothesis of extinction events by a cyclic process is the 26M to 30M year cycle in extinctions proposed by Raup and Sepkoski (1986). More recently, Rohde and Muller (2005) have suggested that biodiversity fluctuates primarily on 62 ± 3 million year cycles.
It is difficult to evaluate the validity of such claims except through reduction to statistical arguments about how plausible or implausible it is for the observed data to exhibit a particular pattern, as the causes of most extinction events are still too uncertain to attribute to them any specific cause let alone a recurring one. Much early work in this area also suffered from the poor accuracy of geological dating, where errors often exceed 10M years. However, improvements in radiometric dating have reduced the scale of uncertainty to at most 4M years — theoretically adequate for studying these processes.
While the statistics alone have been judged as sufficiently compelling to warrant publication, it is important to consider processes that might be responsible for a cyclic pattern of extinctions and future work may focus on trying to find evidence of such processes.
The physicist Richard A. Muller has produced a number of speculative hypotheses for the regularity of mass extinctions. One is that the extinction cycle could be caused by the orbit of a hypothetical companion star dubbed Nemesis that periodically disturbs the Oort cloud, sending storms of large asteroids and comets towards the Solar System.
Muller has also speculated the periodicity of mass extinctions may be related to the solar system's oscillation through the plane of our Milky Way galaxy as it rotates around the galactic centre, with a number of possible hypothesized effects including gravitationally-induced comet showers or periods of intense radiation as the solar system hits the galactic shock wave. 
It has also been suggested that extinction events correlate to the passage of the solar system through the spiral arms of the Milky Way. The earth passes through all four arms every 700 million years, and there is some evidence to suggest a cyclicity of extraterrestrial activity back to 2 billion years ago. and has been elevated above the background level since the Pleistocene. It is feared that 50% of species could be extinct by the end of the 21st century. Factors such as deforestation, habitat destruction, hunting, the introduction of non-native species, pollution and climate change have reduced biodiversity profoundly.
Other hypotheses are that geological instabilities allow heat to periodically build up deep in the Earth, which is then released through mantle plumes, periods of major volcanism and active plate tectonics.
See main article: Holocene extinction event.
The current rate of extinction based on statistical modelling of populations of species for which data is easily obtainable, is estimated to be 10 to 100 times the usual background level,