Speciation, Phylogeny and the History of Life

Natural selection explains how a population changes, but it does not by itself explain why the living world is divided into millions of distinct kinds of organisms rather than one continuous blur. This chapter follows evolution from the split of one population into two, to the branching trees that record those splits, and finally to the four billion year history in which life arose, transformed the planet and survived five catastrophic extinctions. Understanding these processes lets you read a phylogenetic tree correctly, judge claims about where a new disease or crop came from, and see why biodiversity lost today will not quickly return.

Section 1Species concepts and reproductive isolation

Microevolution and macroevolution

The previous chapter treated evolution as change in allele frequencies within a population from one generation to the next. That small scale change is microevolution. Over long spans of time, accumulated microevolution produces macroevolution: the origin of new species, new groups such as flowering plants or mammals, and new structures such as feathers or seeds. The bridge between the two scales is speciation, the process by which one ancestral species gives rise to two or more descendant species. Charles Darwin titled his book On the Origin of Species, yet the question of exactly how one species splits into two was one he could only partly answer. Modern biology answers it by focusing on how gene flow between populations stops.

What is a species?

The word species is Latin for "kind" or "appearance," and for most of history species were defined by how organisms looked. Biologists now recognize several species concepts, each useful for a different purpose. No single definition works for every organism, and that is not a failure of biology; it reflects the fact that speciation is a gradual process, so some populations are caught midway.

The biological species concept, developed by Ernst Mayr, defines a species as a group of populations whose members can interbreed in nature and produce viable, fertile offspring, but cannot do so with members of other such groups. The key idea is reproductive isolation: a species is a gene pool that is sealed off from others. Western meadowlarks and eastern meadowlarks look nearly identical, yet their songs differ and they rarely interbreed where their ranges overlap, so they are separate biological species. Conversely, humans across the world vary greatly in appearance but form a single interbreeding species. This concept has the advantage of focusing on the mechanism that keeps species distinct, namely the absence of gene flow.

The biological species concept has real limits. It cannot be applied to fossils, because we cannot test whether extinct organisms interbred. It does not apply to organisms that reproduce asexually, such as bacteria, many fungi and some lizards and plants, because there is no interbreeding to measure. It is hard to apply to populations that never meet in nature. And gene flow between species that are clearly distinct is more common than once thought: grizzly bears and polar bears, for example, can produce fertile hybrids.

The morphological species concept defines species by body shape and other structural features. It is the oldest approach and the one used in practice for most named species, and it works for fossils and asexual organisms. Its weakness is subjectivity: deciding how much difference counts as a new species is a judgment call, and it can miss cryptic species that look alike but do not interbreed, or lump together very different looking males and females of one species.

The ecological species concept defines a species by its niche, its role in the environment and the set of resources and conditions it uses. Two populations that look similar but occupy different niches are treated as different species. This concept handles asexual organisms well and highlights the role of disruptive selection in creating species.

The phylogenetic species concept defines a species as the smallest group of individuals that share a common ancestor and are distinguished by a unique set of shared traits, forming one tip on the tree of life. With DNA sequencing, this concept has become widely used, especially in microbiology and conservation. Its drawback is that it tends to split groups into many species, since almost any isolated population acquires at least one unique genetic marker.

Reproductive isolation

Under the biological species concept, the defining feature of a species is that its gene pool is isolated. The biological factors that prevent members of different species from producing viable, fertile offspring are called reproductive barriers or isolating mechanisms. They are classified by when they act relative to the formation of a zygote, the fertilized egg. Prezygotic barriers act before fertilization: they prevent mating or block fertilization if mating is attempted. Postzygotic barriers act after a hybrid zygote has formed: they prevent the hybrid from developing into a viable, fertile adult. Most species pairs are separated by several barriers at once.

Prezygotic barriers

• Habitat isolation. Two species living in the same area occupy different habitats and rarely encounter each other. Two species of garter snake in the same region may live one mainly in water and the other mainly on land.
• Temporal isolation. Species breed at different times of day, different seasons or different years. Eastern and western spotted skunks overlap geographically, but one mates in late summer and the other in late winter. Many plants in one meadow flower in different weeks.
• Behavioral isolation. Courtship rituals, songs, colors and pheromones that attract mates are species specific. Blue footed boobies mate only after a high stepping display that shows off the male's feet; fireflies of different species use distinct flash patterns.
• Mechanical isolation. Mating is attempted but morphological differences prevent it. Snails whose shells spiral in opposite directions cannot align their genital openings. Many flowers are shaped to fit only one kind of pollinator.
• Gametic isolation. Sperm of one species cannot fertilize the eggs of another. This is especially important in organisms with external fertilization, such as sea urchins, which release gametes into the water: species specific bindin proteins on sperm must match receptors on the egg surface.

Postzygotic barriers

• Reduced hybrid viability. Genes of the two parent species interact poorly, so the hybrid fails to develop or is frail. Hybrids between some species of salamander often die before reaching maturity.
• Reduced hybrid fertility. The hybrid is vigorous but sterile. If the parent species have different chromosome numbers or structures, homologous chromosomes cannot pair properly in meiosis, so viable gametes are not produced. The mule (horse, 64 chromosomes; donkey, 62; mule, 63) is the classic case.
• Hybrid breakdown. First generation hybrids are viable and fertile, but when they mate with each other or with a parent species, offspring of the next generation are weak or sterile. Some strains of cultivated rice show this pattern.

Why do prezygotic barriers matter so much? Producing a hybrid that dies or is sterile wastes the parents' energy, gametes and time. If hybrids are less fit, natural selection will favor individuals that avoid mating with the other species in the first place. This is the logic of reinforcement, discussed in the next section. As a rule, prezygotic barriers are more "efficient" because they stop wasted reproductive effort earliest.

Section 2Allopatric and sympatric speciation

The central problem: stopping gene flow

Gene flow acts like glue. As long as individuals move between populations and interbreed, alleles are shared and the populations stay similar. For two populations to become separate species, gene flow between them must be reduced enough that mutation, natural selection and genetic drift can push their gene pools in different directions. Once enough differences build up, reproductive barriers arise as a byproduct, and the populations remain distinct even if they later meet. Biologists classify the modes of speciation by how gene flow is interrupted: by geography (allopatric, from Greek words meaning "other country") or within a shared area (sympatric, "same country").

Allopatric speciation

In allopatric speciation, a population is divided into geographically isolated subpopulations. The separation can happen in two broad ways. In vicariance, a barrier arises that splits a continuous population: a mountain range rises, a river changes course, a glacier advances, a land bridge floods or a continent rifts apart. In dispersal, a few individuals colonize a new, distant area, such as an island, and the founding population is isolated from the source. Dispersal often involves a founder effect, so genetic drift can be strong from the very start.

Whether a barrier isolates populations depends on the organism. A highway may block gene flow for small rodents or snails but not for birds or wind dispersed pollen. A canyon a few kilometers wide separates populations of ground squirrels on the north and south rims of the Grand Canyon, yet birds fly across it easily.

Once separated, the isolated populations diverge for several reasons. Their environments differ, so natural selection favors different traits. Different mutations arise in each by chance. Genetic drift, especially in small populations, changes allele frequencies randomly. Sexual selection may favor different mating signals. Over many generations these changes produce reproductive barriers, not because selection "aims" at isolation, but because divergent genomes and behaviors are incompatible as a side effect.

Classic evidence for allopatric speciation includes the snapping shrimp of the genus Alpheus. When the Isthmus of Panama closed about 3 million years ago, it divided populations into Atlantic and Pacific groups. Today there are 15 pairs of sister species, one member on each side of the isthmus, and members of each pair are less likely to mate with each other than with their own side. Pairs from deeper water, which were separated earlier as the isthmus rose, show greater genetic divergence than pairs from shallow water separated last. Similar patterns appear in island groups: the many finch species of the Galápagos and the hundreds of fruit fly species of Hawaii descend from small founding populations that spread from island to island.

Sympatric speciation

In sympatric speciation, a new species arises within the geographic range of its parent population, without physical separation. This is less common than allopatric speciation because gene flow is hard to stop when individuals share an area, but it can happen when something reduces gene flow quickly. Three mechanisms are well documented.

Polyploidy is the presence of more than two complete sets of chromosomes. It can produce a new species in a single generation and is especially common in plants. In autopolyploidy, an individual has more than two chromosome sets all derived from one species. For example, a failure of cell division may turn a diploid cell (2n) into a tetraploid (4n). A tetraploid plant can self fertilize or mate with other tetraploids, but if it mates with a diploid, the triploid offspring (3n) have unpaired chromosomes during meiosis and are sterile. The tetraploids are therefore instantly isolated. In allopolyploidy, two different species interbreed. The hybrid is usually sterile because its two chromosome sets are not homologous and cannot pair in meiosis. But if a later error doubles the hybrid's chromosome number, every chromosome now has a partner, meiosis proceeds and the new allopolyploid is fertile with its own kind yet isolated from both parents. Estimates suggest that a large share of flowering plant species, perhaps 25 to 40 percent or more, arose through polyploidy at some point in their history. Many crops are allopolyploids, including bread wheat (a hexaploid with chromosome sets from three ancestral species), oats, cotton, potatoes, coffee and canola.

Habitat differentiation can also produce sympatric speciation when a subpopulation begins exploiting a resource or habitat not used by the parent population. The North American apple maggot fly, Rhagoletis pomonella, originally laid its eggs on native hawthorn fruits. After apples were introduced roughly two centuries ago, some flies began using apple trees. Because flies tend to mate on the type of fruit they grew up in, and because apples ripen a few weeks earlier than hawthorns, the apple and hawthorn populations have become partly isolated by both habitat and timing. Natural selection has also adjusted the flies' development to each fruit's schedule. These two host races show measurable genetic differences and are considered a case of speciation in progress.

Sexual selection can drive sympatric speciation when females in one population develop preferences for different male traits. In Lake Victoria in East Africa, hundreds of cichlid fish species evolved within a single lake. In clear water, females of two closely related species choose mates by color: one prefers males with blue backs, the other males with red backs. When researchers placed the fish under monochromatic orange light, which hides the color difference, females mated with males of either species and produced viable, fertile hybrids. This shows that mate choice by color was the main barrier. Unfortunately, increasing murkiness of the lake from pollution has weakened this barrier, and some species pairs are merging.

Biologists also describe parapatric speciation, an intermediate case in which populations occupy adjacent areas with a narrow zone of contact. Gene flow continues at the boundary, but strong selection on each side, for example on grasses growing over toxic mine tailings next to normal soil, can still drive divergence.

Hybrid zones

When populations that have partly diverged come back into contact, they may interbreed in a hybrid zone, a region where members of different species meet, mate and produce at least some offspring of mixed ancestry. Hybrid zones are natural laboratories for watching speciation. Fire bellied and yellow bellied toads in Europe, for example, meet along a narrow band stretching hundreds of kilometers; hybrids there have more skeletal abnormalities and lower survival, so alleles from one species rarely penetrate far into the range of the other.

Over time, a hybrid zone can follow one of three paths:

• Reinforcement. If hybrids are less fit than either parent, natural selection strengthens prezygotic barriers, reducing hybrid formation. The species become more distinct where they overlap. Evidence comes from pied and collared flycatchers in Europe: males of the two species look most different in areas where both species live together and more similar where they live apart, the pattern predicted if selection has reinforced mate recognition.
• Fusion. If hybrids are as fit as the parents and barriers are weak, gene flow can increase until the two gene pools merge into one. The Lake Victoria cichlids in murky water illustrate this collapse.
• Stability. Hybrids continue to be produced at a steady rate. This can happen if hybrids are well suited to the zone's particular environment or if steady gene flow from outside the zone balances selection against hybrids. Many hybrid zones have been stable for thousands of years.

The tempo of speciation

The fossil record shows two broad patterns. In many lineages, a species appears suddenly in the rock layers, persists essentially unchanged for millions of years and then disappears. Paleontologists Niles Eldredge and Stephen Jay Gould called this pattern punctuated equilibrium: long periods of stasis punctuated by relatively brief periods of change. Other lineages show gradualism, in which change accumulates steadily and new species emerge slowly. "Brief" in geologic terms can still mean tens of thousands of years, short compared with a species lifetime of several million years but long enough for ordinary natural selection to produce a new form. Studies of speciation intervals in living groups show a wide range: some species pairs separated in a few thousand years (such as cichlids in young lakes), while others took millions of years. On average, the time from initial divergence to complete reproductive isolation is often estimated at around a few million years.

Section 3Phylogenetic trees and cladistics

Phylogeny and classification

A phylogeny is the evolutionary history of a species or group of species. Because every living species descends from earlier species, the history of life can be represented as a branching tree. Reconstructing that tree is the work of systematics, the discipline that classifies organisms and determines their evolutionary relationships using evidence from fossils, anatomy, development and, increasingly, DNA and protein sequences.

Formal naming follows the system introduced by Carolus Linnaeus in the 1700s. Each species receives a two part binomial: the genus name, capitalized, followed by the specific epithet, both italicized. Humans are Homo sapiens. Species are grouped into a nested hierarchy of increasingly inclusive categories: domain, kingdom, phylum, class, order, family, genus, species. Each named unit at any level is a taxon (plural taxa). For the domestic cat, the sequence runs Eukarya, Animalia, Chordata, Mammalia, Carnivora, Felidae, Felis, Felis catus. The ranks above species are human conventions: a genus of beetles is not equivalent in age or diversity to a genus of mammals. For this reason many biologists now prefer to name clades defined by ancestry rather than to assign ranks.

Reading a phylogenetic tree

A phylogenetic tree is a diagram of hypothesized relationships. Learning its parts prevents most errors in interpretation:

• The root represents the most recent ancestor shared by all taxa on the tree.
• Each branch point, or node, represents the divergence of two or more lineages from a common ancestor.
• The tips represent the taxa being compared, usually living species or groups.
• Sister taxa are groups that share an immediate common ancestor not shared by any other group; they are each other's closest relatives on the tree.
• A basal taxon diverges early in the history of the group and lies on a branch that originates near the root.
• A polytomy is a node from which more than two lineages emerge, signaling that the order of divergence is not yet resolved.

Three rules of interpretation are worth memorizing. First, trees show patterns of descent, not overall similarity: two species can look alike yet be distantly related. Second, unless branch lengths are scaled, the length of a branch says nothing about time or the amount of change. Third, no living taxon is the ancestor of another living taxon. Humans did not evolve from chimpanzees; both descend from a common ancestor that lived roughly 6 to 7 million years ago. The order of tips along the top of a tree is also meaningless: branches can rotate around any node without changing the relationships shown, just as a hanging mobile can spin.

Homology and analogy

Phylogenies are built from characters, heritable features that vary among taxa, such as the presence of a backbone or the base at a given position in a gene. Useful characters must be homologous: similar because they were inherited from a common ancestor. The forelimbs of whales, bats and humans contain the same set of bones because they descend from the limb of an ancestral tetrapod. In contrast, analogous structures are similar because of convergent evolution, in which unrelated lineages face similar selective pressures and evolve similar solutions. Bird wings and insect wings both allow flight but are built from completely different tissues. Similarity due to convergence is called homoplasy, and it can mislead anyone who assumes that resemblance always means relatedness. Generally, the more complex and similar two structures are in detail, the more likely they are to be homologous.

Cladistics

Cladistics is the approach that groups organisms by common descent. Its basic unit is the clade, a group consisting of an ancestral species and all of its descendants. Clades are nested, like boxes within boxes: mammals form a clade within the larger clade of amniotes, which lies within tetrapods, which lies within vertebrates.

Groupings are judged by whether they form a true clade:

• A monophyletic group (a clade) contains an ancestor and all of its descendants. Birds plus all the dinosaurs they descend from form a monophyletic group.
• A paraphyletic group contains an ancestor and some, but not all, of its descendants. The traditional "reptiles," defined to exclude birds, is paraphyletic because birds descend from the same ancestor as crocodiles and lizards.
• A polyphyletic group includes distantly related species but not their most recent common ancestor. Grouping all "warm blooded animals" (birds and mammals) together is polyphyletic, because endothermy evolved independently in the two lineages.

Cladistic analysis distinguishes two kinds of character states. A shared ancestral character originated in an ancestor of the whole group being studied, so it does not help sort members within that group. A backbone is shared ancestral within mammals: every mammal has one, so its presence tells you nothing about which mammals are most closely related. A shared derived character (a synapomorphy) is an evolutionary novelty unique to a particular clade. Hair is a shared derived character of mammals. Whether a character counts as ancestral or derived depends on the level being studied: a backbone is derived when vertebrates are compared with other animals.

To decide which states are ancestral, systematists compare the group of interest, the ingroup, with an outgroup, a species or group known to have diverged before the ingroup lineages. Characters shared by the outgroup and the ingroup are assumed to be ancestral; those found only in some members of the ingroup are derived. In the tree in Figure 9.3, the lancelet serves as the outgroup for the vertebrates.

Choosing among trees: parsimony and likelihood

Even for a few species, many possible trees exist; for 50 species there are more trees than atoms in the observable universe. Systematists use principles to pick the best supported tree. Maximum parsimony chooses the tree that requires the fewest evolutionary changes, following the idea (often called Occam's razor) that the simplest explanation consistent with the evidence should be preferred first. For DNA data, the most parsimonious tree requires the fewest base changes. Maximum likelihood and Bayesian methods use explicit models of how DNA changes over time, for example that transitions between purines are more frequent than transversions, and find the tree most likely to have produced the observed data. A tree is a hypothesis that is revised as new data appear.

Molecular data and molecular clocks

DNA comparisons have transformed systematics because they supply thousands of characters for any organism. Genes evolve at different rates. Ribosomal RNA genes change slowly and are used to study divergences hundreds of millions of years old, including the relationships among the three domains. Mitochondrial DNA changes rapidly and is used to compare closely related species and populations.

A molecular clock estimates the time of evolutionary change from the number of differences between homologous sequences. It rests on the observation that some genes accumulate neutral mutations at a fairly constant average rate. The clock is calibrated by plotting the number of differences against divergence times known from well dated fossils. Molecular clocks must be used with care, because rates can differ among genes and lineages, but they have dated events such as the origin of HIV: by comparing samples from patients, researchers estimated that the HIV-1 M strain spread from chimpanzee viruses into humans in the early twentieth century.

The three domains and the tangled tree

Comparisons of ribosomal RNA in the 1970s, pioneered by Carl Woese, revealed that prokaryotes include two profoundly different lineages. Life is now divided into three domains: Bacteria, Archaea and Eukarya. Archaea share some features with eukaryotes (such as histone proteins and similar RNA polymerases) and are generally considered more closely related to Eukarya than to Bacteria; some analyses place eukaryotes within the Archaea. The base of the tree is complicated by horizontal gene transfer, the movement of genes between genomes by plasmids, viruses and even the fusion of organisms. Because genes have moved sideways between lineages throughout the history of life, the early tree looks less like a single trunk and more like a web or tangled ring.

Section 4The origin of life and the fossil record

The early Earth

Earth formed about 4.6 billion years ago (more precisely about 4.54 billion) from material orbiting the young sun. For its first several hundred million years it was bombarded by debris, and one giant collision probably produced the Moon. As the planet cooled, water vapor released by volcanoes condensed into oceans. The early atmosphere was thick with water vapor, nitrogen, carbon dioxide and smaller amounts of methane, ammonia, hydrogen and hydrogen sulfide, but it had almost no free oxygen. The oldest direct evidence of life, fossil stromatolites (layered mats built by communities of microbes trapping sediment), dates to about 3.5 billion years ago, and chemical traces in older rocks suggest life may be about 3.7 billion years old or more. Life therefore appeared within a few hundred million years of Earth becoming habitable.

From chemistry to life: four stages

No one can observe the origin of life directly, but biologists have identified a plausible sequence of stages, each of which can be tested in the laboratory:

1. Abiotic synthesis of small organic molecules such as amino acids and nitrogenous bases.
2. Joining of these monomers into polymers such as proteins and nucleic acids.
3. Packaging of molecules into protocells, membrane bounded droplets with an internal chemistry different from their surroundings.
4. Origin of self replicating molecules that made inheritance possible.

In the 1920s, A. I. Oparin and J. B. S. Haldane independently proposed that the early atmosphere, lacking oxygen, would allow organic molecules to form and accumulate, since oxygen would otherwise break them down. In 1953, Stanley Miller and Harold Urey tested this idea. They circulated water vapor, hydrogen, methane and ammonia through a sealed apparatus, sending electrical sparks through the gases to mimic lightning. Within a week, the water contained a variety of amino acids and other organic compounds. Although the early atmosphere may have been less reducing than Miller assumed, organic molecules can also form near deep sea hydrothermal vents, particularly alkaline vents where warm, mineral rich water creates natural chemical gradients. Meteorites add further support: some contain amino acids, sugars and nitrogenous bases formed in space, showing that these building blocks arise readily under abiotic conditions.

Monomers can link into polymers without enzymes when dilute solutions drip onto hot sand, clay or rock, which evaporates the water and concentrates the molecules. Clay surfaces bind charged monomers and catalyze their joining; RNA monomers have been shown to form chains of 50 or more nucleotides on clay.

Protocells could form spontaneously because lipids and other amphipathic molecules self assemble into bilayer vesicles in water. In laboratory experiments, montmorillonite clay speeds up vesicle formation and can carry RNA inside. Such vesicles can grow by absorbing more lipid, divide when sheared and maintain an internal environment, showing several properties of life.

The RNA world

Modern cells use DNA to store information and proteins to catalyze reactions, but each depends on the other: DNA cannot be copied without protein enzymes, and proteins cannot be made without information from nucleic acids. Which came first? The discovery of ribozymes, RNA molecules that act as catalysts, suggested an answer. RNA can both store genetic information and catalyze reactions, including, in laboratory evolved ribozymes, the copying of short stretches of RNA. The RNA world hypothesis proposes that early life used RNA for both functions. Natural selection could act on RNA molecules directly: sequences that folded into shapes that copied themselves faster or more accurately would come to dominate. Strong evidence comes from modern cells: the catalytic core of the ribosome, which forms peptide bonds, is made of RNA, a molecular fossil of that earlier world. DNA, more chemically stable, later took over as the main store of information, and proteins, more versatile catalysts, took over most enzymatic work.

The fossil record

A fossil is any preserved remnant or trace of an organism from the past. Most fossils form in sedimentary rock, which builds up as sand and mud settle in layers called strata in seas, lakes, swamps and floodplains. Deeper strata are generally older. Organisms buried quickly in fine sediment may become fossils when minerals gradually replace their hard parts, a process that produces bones and shells of stone. Other fossils are molds and casts, footprints and burrows (trace fossils), thin carbon films of leaves, insects in amber, or, rarely, frozen or dried bodies with soft tissue.

The fossil record is detailed in places but strongly biased. Organisms are more likely to fossilize if they had hard parts, lived in or near water where sediments accumulate, existed for a long time, were abundant and widespread, and died in a place where the rock was later neither destroyed nor deeply buried. Soft bodied organisms, small populations and land animals of eroding uplands are underrepresented. So the absence of a group from the record does not prove it was absent from Earth. Despite these gaps, fossils document many transitions in detail, including the evolution of tetrapods from lobe finned fishes, of whales from land mammals and of mammals from earlier synapsids, in which a series of fossils shows the jaw joint shifting while two of its bones moved into the middle ear.

Dating rocks and fossils

Two complementary approaches give the ages of fossils. Relative dating orders events by their position in the strata: lower layers are older than upper layers, and fossils found in the same layer in different places lived at roughly the same time. Relative dating tells you the sequence of events but not their age in years. Radiometric dating provides absolute ages. It relies on radioactive isotopes, or "parent" isotopes, which decay to "daughter" isotopes at a fixed rate unaffected by temperature or pressure. The rate is expressed as a half-life, the time required for half of the parent isotope in a sample to decay. After one half-life 50 percent remains; after two, 25 percent; after three, 12.5 percent; and so on.

Different isotopes suit different time spans. Carbon-14, with a half-life of 5,730 years, is used to date organic remains up to about 50,000 to 75,000 years old. Living organisms take in carbon from the environment and keep the same ratio of carbon-14 to carbon-12 as the atmosphere; at death, intake stops and the carbon-14 begins to decay. A bone with one quarter of the original carbon-14 ratio is about two half-lives, or 11,460 years, old. For older fossils, geologists date volcanic layers above and below the fossil using isotopes with longer half-lives, such as potassium-40 (half-life about 1.3 billion years, decaying to argon-40) or uranium-238 (about 4.5 billion years, decaying through a series to lead-206). A fossil sandwiched between ash layers dated to 525 and 535 million years is roughly 530 million years old. Sedimentary rock itself is usually not dated directly because its mineral grains come from older rocks.

Section 5Major events and mass extinctions in the history of life

The geologic time scale

Geologists divide Earth's history into a hierarchy of eons, eras and periods, with boundaries usually set by major changes in the fossils found in the rocks. The Hadean (4.6 to 4.0 billion years ago) and Archean (4.0 to 2.5 billion) eons saw the formation of the planet and the first prokaryotes. The Proterozoic (2.5 billion to 541 million years ago) saw oxygen accumulate and eukaryotes arise. The Phanerozoic ("visible life"), the last 541 million years, contains three eras: the Paleozoic ("old life"), Mesozoic ("middle life," the age of reptiles) and Cenozoic ("recent life," the age of mammals).

Milestones of early life

For more than a billion years after the first stromatolites, all life was prokaryotic. The most transformative event of this time was oxygenic photosynthesis by cyanobacteria, which split water and released O2. At first the oxygen reacted with dissolved iron and precipitated as the red banded iron formations found in rocks of this age. Once these sinks were filled, oxygen began to accumulate in the atmosphere during the Great Oxidation Event, about 2.4 billion years ago. Oxygen was toxic to many anaerobic prokaryotes, which went extinct or retreated to anaerobic habitats, but it also made aerobic respiration, which yields far more ATP, possible, and eventually it formed an ozone layer that shielded the land from ultraviolet light.

The oldest widely accepted fossils of eukaryotic cells are about 1.8 billion years old. Eukaryotes arose through endosymbiosis: an archaeal host cell engulfed an aerobic bacterium that became the mitochondrion, and later a lineage engulfed a cyanobacterium that became the chloroplast (Chapter 10). Multicellular eukaryotes, such as small red algae, appear by about 1.2 billion years ago. A series of severe "snowball Earth" glaciations from about 750 to 580 million years ago may have limited the diversity of early multicellular life. Larger, soft bodied animals of the Ediacaran biota appeared about 635 to 541 million years ago.

The Cambrian explosion and the move onto land

The Cambrian explosion, beginning about 535 to 525 million years ago, is the sudden appearance in the fossil record of most of the major animal phyla living today, including the first animals with hard, mineralized skeletons, predators with claws and prey with spines and armor. Molecular clocks suggest that many animal lineages originated somewhat earlier, so the "explosion" partly reflects the evolution of hard parts that fossilize well, along with rising oxygen and predator prey arms races. Life then spread onto land: plants, fungi and arthropods by about 500 to 470 million years ago, forests and the first four limbed vertebrates (tetrapods) by about 365 million years ago.

Continental drift

Earth's crust is broken into tectonic plates that move a few centimeters a year over the hot mantle. About 250 million years ago, near the end of the Paleozoic, plate movements brought nearly all land together into the supercontinent Pangaea. Its formation destroyed much shallow coastal habitat, made the interior dry and changed ocean currents, contributing to extinctions. When Pangaea began to break up about 200 million years ago, populations on separating landmasses evolved in isolation, a giant case of allopatric speciation. This explains many patterns in biogeography: matching fossils of the same freshwater reptile in South America and Africa, and the distinctive marsupials of Australia, which drifted away before placental mammals spread there.

Mass extinctions

Extinction is normal: the vast majority of species that ever lived are now extinct, most during the steady "background" rate of extinction. A mass extinction is different: a large proportion of species, often more than half, disappears within a geologically short interval because of a global environmental change. The fossil record shows five mass extinctions, the "big five." Two are especially well studied:

• The Permian extinction, at the boundary between the Paleozoic and Mesozoic eras about 252 million years ago, was the largest. It killed about 96 percent of marine species and a large fraction of land species, and it occurred in less than about 500,000 years. It coincided with enormous volcanic eruptions in what is now Siberia, which released carbon dioxide, warmed the climate, acidified and deoxygenated the oceans and may have favored bacteria that produced toxic hydrogen sulfide.
• The Cretaceous extinction, about 66 million years ago, eliminated about three quarters of species, including all nonavian dinosaurs. A thin clay layer at the boundary, rich in iridium (an element rare on Earth but common in meteorites), and the buried 180 kilometer Chicxulub crater off the Yucatán Peninsula indicate that an asteroid struck Earth. Dust and soot blocked sunlight, collapsing food webs for months to years.

Consequences: recovery and adaptive radiation

Mass extinctions change the course of evolution because they remove whole lineages regardless of how well adapted they were to normal conditions. Recovery of diversity typically takes 5 to 10 million years or more. Survivors find empty ecological roles and undergo adaptive radiation, the evolution of many diversely adapted species from a common ancestor. After the Cretaceous extinction, mammals, which had been mostly small and nocturnal for 140 million years, radiated into whales, bats, hoofed animals and primates. Adaptive radiations also follow the evolution of key innovations (such as wings or flowers) and colonization of new areas, such as the Hawaiian silverswords descended from a single California tarweed ancestor.

Words worth knowing