The Long March Onto Land

The Question

How did life become visible?

For the first three and a half billion years of its existence on Earth, life was almost entirely microscopic. There were stromatolite mats, and toward the end there were filamentous algae and a handful of decimetre-scale aggregates, but to a casual observer walking along the shoreline of any continent in 2026 these would not register as life at all. The sea was busy with cells, but the world above the tideline was bare rock. By the time the period covered in this entry ends, that situation has been transformed beyond recognition. There are animals with eyes and guts and limbs. There are forests. There are creatures that lay eggs in dry soil and walk away. The change took roughly four hundred million years, and the events that drove it are mostly preserved well enough in the rock record that we can describe them in some detail.

This is the story of how life on Earth left the microscope and stepped onto the land.

Snowball Earth

The transition begins, oddly, with the planet freezing.

At intervals during the period now called the Cryogenian — running from roughly 720 to 635 million years ago — Earth fell into the most severe ice ages in its geological record. Glacial sediments of this age have been recovered on every modern continent, including localities that were unambiguously near the equator at the time of deposition. The simplest interpretation, advanced in detail in the late 1990s and progressively refined since, is that ice sheets reached all the way from the poles to the tropics, and that the surface of the oceans froze across most or all of the planet. The hypothesis is now widely referred to as Snowball Earth ^[1].

There were two principal episodes. The Sturtian glaciation, the older and longer of the two, began at around 717 million years ago and persisted for some 57 million years before retreating at around 660 million years ago. The Marinoan glaciation came roughly 15 million years later and ended at around 635 million years ago, with recent radiometric work constraining its fully-glaciated phase to perhaps four million years ^[2]. Both episodes were accompanied by the kind of carbon-isotope excursions in ancient sediments that are characteristic of severe and prolonged disruptions of the global carbon cycle.

The mechanism that triggered them is not yet completely understood. The leading account ties the onset of the Sturtian event to the break-up of the supercontinent Rodinia and a corresponding drop in atmospheric carbon dioxide as freshly exposed silicate rock weathered and consumed CO₂ from the air. Once enough ice had formed at high latitudes, the increased reflectance of the Earth's surface fed a runaway cooling: more ice meant more reflected sunlight meant more cooling meant more ice. The runaway eventually stopped only when ice covered enough of the surface that very little volcanic CO₂ was being weathered out of the atmosphere; over millions of years, volcanic outgassing built atmospheric CO₂ back up to the point where the greenhouse effect overwhelmed the planet's reflectance and the ice retreated, abruptly, into a hot post-glacial world ^[1].

What matters for this entry is not the climatology but the timing. The Cryogenian glaciations bracket the period immediately preceding the appearance of the first macroscopic animals. Whether the freezing acted as a filter that selected for new biological strategies, or supplied chemical conditions favourable to larger bodies (oxygenation pulses, nutrient flushes from glacial weathering have both been proposed), or simply preceded the new biology by coincidence, is not yet settled. But the next chapter of life follows immediately on the heels of the last great snowball.

The First Large Bodies

The retreat of the Marinoan ice sheets opens the Ediacaran period, running from 635 to 538.8 million years ago. Within this interval, for the first time in Earth's history, there are organisms in the rock record large enough that an unaided human eye would notice them.

The earliest of these are preserved as impressions on ancient sea-floor sediments at Mistaken Point in Newfoundland, in deep-water deposits dating to roughly 574 million years ago. The forms are unlike anything alive today. Many are rangeomorphs — frond-shaped organisms whose surface is covered with a self-similar fractal branching pattern that repeats across at least four levels of scale. The largest examples reached up to about two metres tall. They were anchored to the sea floor by a basal disc and presumably absorbed nutrients directly from seawater, having no mouth, no gut, and no detectable internal organs of any kind. Charnia masoni, first described from rocks of similar age in Charnwood Forest in England, is the type specimen of the group ^[3].

Slightly younger Ediacaran assemblages, dated between roughly 560 and 550 million years ago and preserved in localities including the White Sea region of Russia, the Flinders Ranges of Australia, and Namibia, contain a wider range of forms. Among these are Dickinsonia, an oval, ribbed, mat-like organism reaching as much as a metre in length, and Kimberella, a smaller bilaterally symmetrical creature that left scrape marks on the sea floor consistent with a muscular foot used to graze the microbial mats it lived among. There are also tubular and stalked forms, frond-like forms, and a small number of organisms that may anticipate later sponge or cnidarian body plans.

The biological affinities of the Ediacaran organisms remain genuinely contested in 2026. Charnia and the rangeomorphs do not match any modern phylum cleanly and may belong to an entirely extinct branch of multicellular life. Dickinsonia yielded animal-specific cholesteroid biomarkers in a 2018 study of preserved organic residues, and on that basis is now widely treated as an early animal, although what kind of animal remains unclear ^[4]. Kimberella is the most plausible early bilaterian — an organism with a clearly defined front and back, top and bottom — and is the best candidate for an Ediacaran member of a lineage with descendants alive today. The most cautious framing is that the Ediacaran biota records the first body-plan-scale experiment in macroscopic multicellular life, that some of its members were animals, and that some were almost certainly not.

By the close of the period, around 540 million years ago, most of the distinctive Ediacaran forms had disappeared from the fossil record. Whether they were actively eliminated by the appearance of the new mobile, predatory animals that came next, or whether they declined for environmental reasons that the rocks have not yet revealed, is one of the active questions of early-animal palaeontology.

The Cambrian Explosion

Beginning at the boundary now formally placed at 538.8 million years ago ^[5], the rock record undergoes a transformation so abrupt and so dramatic that it has been recognised as a distinct event for nearly two centuries. The Cambrian Explosion is the name now given to the relatively brief geological interval — perhaps twenty to twenty-five million years — during which most of the major body plans of the modern animal kingdom appear in the fossil record.

The animals of the Cambrian had hard parts: shells, exoskeletons, plates, spicules, and toothed feeding apparatuses preserved as small phosphatic fossils called small shelly fossils. They had appendages. They had compound eyes. They had guts running from a mouth at one end to an anus at the other. They had a recognisable predator–prey ecology, with armour, claws, and burrowing as evident defensive and offensive responses to one another. Within this twenty-five-million-year stretch, the rocks register the first clear appearance of arthropods, molluscs, brachiopods, echinoderms, chordates, annelid worms, priapulid worms, and several other phyla. Effectively every major animal body plan present on Earth in 2026 — and several that have since gone extinct — first shows up here.

The exceptional preservation of soft tissues in two world-class fossil deposits has shaped our picture of the Cambrian fauna more than any other source. The older of the two, the Chengjiang biota, in what is now Yunnan Province in southwestern China, dates to roughly 518 million years ago and preserves more than two hundred described species of soft-bodied animals at extraordinary resolution, including digestive tracts, eyes, and nerve cords. The younger and more famous, the Burgess Shale in British Columbia in Canada, dates to roughly 508 million years ago and contains around 165 species, many shared with Chengjiang and many unique ^[6]. Both are lagerstätten — exceptional fossil deposits in which conditions have permitted the preservation of soft-bodied organisms that the ordinary fossil record, dominated by hard parts, cannot record. Without them, our picture of the Cambrian would be a fauna of shells and sclerites; with them, it is a fauna of recognisable, complete, often peculiar animals.

The Cambrian creatures preserved at Chengjiang and Burgess include the long-tailed lace-crab Marrella, the spiny worm Hallucigenia, the five-eyed proboscid Opabinia, the swimming disc-shaped predator Peytoia, and Anomalocaris, an apex predator reaching up to a metre in length, with grasping frontal appendages and a circular tooth-ringed mouth. Several lineages defied early classification entirely and have only been securely placed in the modern animal tree in the last few decades, often after revisions to which way up they were drawn or which end was the head. The general pattern, however, is now clear: the Cambrian was not a chaos of radically alien designs; it was the first appearance, in the rock record, of the basic body plans that have organised animal life ever since.

What caused the explosion is debated. Contributing factors that are widely accepted in 2026 include a sustained rise in ocean and atmospheric oxygen at the close of the Ediacaran, the genetic toolkit of bilaterian development (the Hox genes and their regulatory partners) becoming fully assembled, and ecological feedbacks once predation and burrowing took hold. None of these is the whole answer. The Cambrian Explosion is not a sudden invention of complex animals from nothing — the molecular clocks, calibrated against the protein and DNA sequences of living animals, place the divergence of the major animal lineages well back into the Ediacaran or even earlier. What is sudden is the joint appearance of those lineages in the fossil record, in body forms large enough and hard enough to preserve. The explosion is, to a substantial degree, a fossilisation event.

The First Vertebrates

Among the Chengjiang animals, two small fish-shaped creatures occupy a position of particular interest. Myllokunmingia fengjiaoa and Haikouichthys ercaicunensis, both described from Yunnan in the late 1990s, are the oldest known fossils widely interpreted as early members of the vertebrate stem lineage ^[7]. They are small — three centimetres long, more or less — and lack jaws, paired fins, and almost everything else a modern fish would be expected to possess. But they have a notochord, segmented muscle blocks, gill arches, a head with eyes, and the beginnings of a vertebral column. They are recognisably the kind of animal from which all later vertebrates, including the writer and reader of this entry, descend.

The Cambrian vertebrates were not impressive predators. They were slim, sluggish, jawless filter-feeders or scavengers, in oceans dominated by larger arthropods and stranger forms. For something like a hundred million years after their first appearance, the vertebrate lineage would remain a minor part of marine ecosystems. But the basic body plan — head end, tail end, dorsal nerve cord, ventral gut, segmented muscles, internal skeleton — was now in place.

The next two hundred million years would see vertebrates acquire jaws (sometime in the Ordovician or early Silurian, more than four hundred million years ago); paired fins; bone; the divergence of the cartilaginous fish (sharks, rays, chimaeras) from the bony fish (almost everything else); and, within the bony fish, the divergence of the ray-finned lineage that produced the great majority of modern fish from the lobe-finned lineage that produced the tetrapods. The lobe-finned lineage is the one this entry follows onto land.

Plants Go Ashore

While the vertebrates were getting started in the seas, a different and even more consequential migration was beginning along the shorelines of the early continents.

The land surface of the early Cambrian world was largely lifeless above the tideline, beyond a thin film of cyanobacterial and lichen-like crust that may have already covered moist rock and soil. The ozone layer, in place since the consequences of the Great Oxygenation Event a billion and a half years earlier, made the surface chemically habitable. What was missing was a lineage that had solved the structural problems of living above water: how to keep cells from drying out in air, how to extract nutrients from rock and soil rather than seawater, how to reproduce when the gametes can no longer simply swim through their surroundings, and how to stand up against gravity without the buoyant support that water provides.

The lineage that solved these problems came from a freshwater group of green algae closely related to the modern charophytes. Molecular phylogenies place the origin of land plants firmly within the charophyte algae, and the first traces of plants on land are not body fossils but microscopic cryptospores — durable, walled spore-like structures very similar to the spores of modern bryophytes. Cryptospores appear in the rock record from at least the Middle Ordovician onward, around 470 million years ago and possibly earlier, scattered through marine sediments to which they had been washed from nearby coasts ^[8]. The plants that produced them have not been preserved as body fossils — they were probably small, soft, and easily decomposed — but the spores demonstrate that something resembling land plants was producing wind-borne or water-borne reproductive structures by the middle of the Ordovician.

The earliest body fossils of land plants come tens of millions of years later. The first widely accepted fossils belong to a group of small, leafless, branching organisms with a central vascular strand for moving water and dissolved nutrients. Cooksonia barrandei, described in 2018 from rocks in the Czech Republic dated to about 432 million years ago, is the oldest widely accepted megafossil of a vascular land plant in 2026 ^[9]. Cooksonia plants were a few centimetres tall, lacked leaves, lacked roots in the modern sense (they probably grew from a rhizome), and bore terminal spore-producing capsules at the tips of their branches. They were the first members of the vascular plants — the lineage that includes every modern fern, conifer, and flowering plant — and their innovation was the system of internal tubes that lets a plant move water from a wet substrate up to tissues exposed to dry air.

What followed in the Devonian, between roughly 420 and 360 million years ago, was a complete transformation of the continental surface. Vascular plants developed roots, leaves, woody tissue, and seeds. Heights increased from centimetres to metres to tens of metres. By the late Devonian, the progymnosperm Archaeopteris had grown into the planet's first true forests, with continuous canopies, deep root systems, and the modified soil chemistry that established forests still produce. By the Carboniferous — roughly 360 to 300 million years ago — vast lowland swamp forests of giant lycopsids, horsetails, and tree ferns covered the equatorial continents. The carbon those forests pulled from the atmosphere and locked away as undecomposed plant matter eventually became, after burial and compression, the coal seams that powered the human Industrial Revolution. The plants of the Carboniferous, in burying their dead, set aside a chemical fuel that human civilisations would use to remake themselves more than three hundred million years later.

The First Animals On Land

Animals followed plants ashore. The first to make the move successfully were arthropods — the lineage of segmented, jointed-limbed, externally armoured creatures that already dominated the Cambrian seas.

The earliest direct fossil evidence of an animal living on land in air is Pneumodesmus newmani, a small millipede known from a single specimen recovered from rocks at Stonehaven in Aberdeenshire, Scotland. The specimen was originally assigned a late Silurian age of about 428 million years ago, but a subsequent redating based on zircon and palynological evidence from the host rock argued for an early Devonian age of around 414 million years ago instead. Both ages are still defended in the 2026 literature, and the precise age of Pneumodesmus remains an open question ^[10]. The crucial detail is not the size of the animal but the spiracles preserved in its cuticle: small openings into a tracheal system, the kind of gas-exchange apparatus that only works in air. Pneumodesmus was breathing air, on land, more than four hundred million years ago.

Tracks and burrows preserved in older rocks suggest that arthropods may have been making brief excursions onto land considerably earlier, perhaps as far back as the late Cambrian or Ordovician, but unambiguous body fossils of fully terrestrial arthropods only become common in the Silurian and early Devonian. By the late Devonian, terrestrial ecosystems include not only millipedes but mites, springtails, the first scorpions, and the first wingless insects. By the Carboniferous, with atmospheric oxygen at unusually high levels and forests providing both cover and food, some of these arthropods reached sizes never matched since: dragonfly-like insects with wingspans approaching three-quarters of a metre, and millipede-like Arthropleura growing to lengths of two and a half metres on the forest floor. Arthropods had the land mostly to themselves for several tens of millions of years before vertebrates joined them.

From Fins To Limbs

The vertebrate move onto land is the best-documented major transition in the fossil record. It happened in the late Devonian, between roughly 385 and 360 million years ago, in a lineage of lobe-finned bony fish that lived in shallow, warm, often oxygen-poor freshwater habitats edged by the new vascular-plant vegetation.

The intuitive case for the move starts with the limitations of being a fish in those waters. Warm water carries less dissolved oxygen than cold water; the swampy, vegetation-choked margins of a Devonian river system would have been particularly oxygen-starved. A fish that could supplement its gills by gulping air at the surface had a clear short-term advantage. A fish that could prop itself up on its fins, in shallow water, to keep its head clear and its body stable while breathing, had a further advantage. From there, the same body plan that lets a lobe-finned fish lever itself partly out of the water also lets it move, briefly and clumsily, across very short stretches of mud between pools. Selection on each step of that capacity could readily proceed, in any population in which the adults were already living in shallow, weedy water, without any individual ever having to "intend" to colonise dry land.

A handful of fossil organisms from this stretch are the canonical illustrations of the transition. Tiktaalik roseae, described in 2006 from rocks in the Canadian Arctic dated to around 375 million years ago, is the most famous ^[11]. Tiktaalik is unmistakably a fish — it has scales, gills, and fins — but its fins contain a recognisable wrist joint with the equivalent of upper arm, forearm, and digit-supporting bones, all in roughly the arrangement found in later limbs. It has a broad flat head with eyes on top, a mobile neck (a feature absent in standard fish, which have their skull and shoulder girdle fused), and ribs strong enough to support the body's weight against gravity. Subsequent reconstructions of the axial skeleton, published in 2024, have shown that its rear ribs were attached to a robust pelvis, supporting the load-bearing role of the hind half of the body in shallow-water propping ^[12]. Tiktaalik is not the ancestor of the tetrapods — no specific fossil ever fills that role uniquely — but it is an extraordinarily clear example of the kind of animal from which the tetrapods came.

A few million years later, in the latest Devonian, fully limbed animals appear in the rock record. The most thoroughly studied are Acanthostega gunnari and Ichthyostega stensioei, both from East Greenland, both dated to around 365 to 370 million years ago ^[13]. Both have four limbs with digits, but neither was a competent walker. Acanthostega, the better understood of the two, had eight digits on each forelimb, a fish-like tail with a fin, internal gills, and a skeleton that suggests it lived almost entirely in water and used its limbs primarily for moving through submerged vegetation rather than across dry ground. Ichthyostega was sturdier, with seven digits on each hindlimb, a more robust ribcage, and stronger limbs, but its limb structure overall suggests it walked on land only with difficulty. The first tetrapods were aquatic animals with legs, not terrestrial animals with vestigial fish features. Their move onto land was incomplete in their own bodies.

The benchmark for the very earliest tetrapod has moved repeatedly in the past two decades. Trackways preserved in tidal-flat sediments at Zachełmie in Poland, described in 2010, have been dated to around 395 million years ago — some twenty million years earlier than Tiktaalik and thirty million years earlier than Acanthostega and Ichthyostega ^[14]. If the tracks were genuinely made by a four-limbed walking vertebrate, then tetrapods (in the sense of vertebrates with feet) must have existed well before any of the famous transitional body fossils. The interpretation of the Zachełmie tracks has been challenged on a number of grounds and remains an active question, although palaeoenvironmental reanalyses published since 2018 have offered support for the original tetrapod identification. The honest summary is that the canonical late-Devonian transitional fossils are a window into the kind of animal involved in the move from water to land, but not necessarily the very first animals to make the move.

What Amphibians Could And Could Not Do

By the early Carboniferous, around 340 million years ago, vertebrates with limbs were established in the freshwater wetlands of the equatorial continents. They had largely solved the structural problems of supporting a body in air — limbs, ribs, lungs, mobile necks — and they had acquired the basic equipment of terrestrial life. They had not solved every problem.

The decisive remaining problem was reproduction. The early tetrapods, like modern amphibians — the surviving lineage that includes frogs, salamanders, and the limbless caecilians — were tied to standing water by their eggs. Amphibian eggs are coated in a soft, gelatinous, permeable membrane. If laid in air, they dry out within hours. If laid underwater, they exchange gases and waste through the membrane and develop normally; the larva that hatches (a tadpole, in modern frogs) typically continues to live in water, breathing through gills, until it metamorphoses into the adult form. An amphibian, however well it walks, must return to water — or at least to permanently moist mud — to breed.

This is a serious limitation. It restricts the entire group to within a short distance of standing freshwater, and it constrains where new species can radiate. The interior of a continent, with its dry ground, its variable rainfall, and its rivers separated by long stretches of arid land, is not accessible to an animal whose offspring must spend the first part of their lives as fish.

The amphibians of the Carboniferous were nevertheless successful. Some grew large — the Eryops-grade animals reached lengths of two metres, with massive ribcages and crocodile-like ecologies along Carboniferous river margins. The group diversified into a variety of forms, including some that have no clear modern analogues. But none of these animals strayed far from water, and the dry interior of the supercontinent assembling at the end of the Palaeozoic remained, for vertebrates, almost empty.

The Egg That Broke The Water Line

The next innovation broke that constraint. Sometime in the early-to-middle Carboniferous, in a lineage of small, lizard-like tetrapods living in the equatorial swamp forests, the egg itself was redesigned.

The new egg surrounded the embryo in a series of internal membranes. The innermost, the amnion, enclosed the embryo in a fluid-filled cavity that functioned as a private water-filled environment, mimicking the conditions in which an amphibian larva would otherwise have to develop. A second membrane, the chorion, lay just inside the eggshell and handled gas exchange between the embryo and the outside air. A third, the allantois, served as a sac for collecting the metabolic waste the embryo produced during development. The whole package was enclosed in a flexible leathery or calcified shell that retained moisture and resisted drying.

This is the amniotic egg. With it, the embryo no longer required external standing water in which to develop. The egg carried its own water along with it. It could be laid on a dry surface — soil, a rock crevice, a layer of leaf litter — and the embryo would develop unattended. The hatchling that eventually broke out of it was already a miniature version of the adult, breathing air with lungs from the moment it left the shell. The larval, water-bound stage of the amphibian life cycle had been engineered out of existence.

The animal that carries this innovation is, by definition, an amniote. The earliest unambiguous amniotes preserved in the fossil record are small, lizard-like animals of the late Carboniferous. Hylonomus lyelli, known from fossils preserved inside the hollow stumps of fossilised lycopsid trees at Joggins on the Bay of Fundy in Nova Scotia, is roughly 312 million years old and has been the canonical earliest unambiguous amniote for several decades ^[15]. (A slightly older Scottish fossil, Westlothiana lizziae, at around 338 million years old, has often been mentioned as a candidate earlier amniote, but its anatomy is now usually interpreted as that of a stem-amniote rather than a true amniote, and its precise placement is unsettled. As with all of these "earliest" claims, the benchmark is a moving target, and recent phylogenetic work has revisited the placement of Hylonomus itself.)

What matters is not the precise identity of the first amniote but the consequence of the trait. Once the amniotic egg existed, the dry interior of the continents was open to vertebrates for the first time. A lineage that could lay its eggs on dry ground could colonise habitats that no amphibian could reach. From this small beginning, in the swamp forests of the late Carboniferous, an entire half of the modern terrestrial vertebrate world would unfold.

What Follows

The four hundred million years covered in this entry took life on Earth from a microscopic, almost entirely aquatic, almost entirely invisible affair to a world of forests, of insects, of fish, of amphibians, and finally of a small lizard-like amniote sheltering in a hollow tree on the coast of what is now Nova Scotia. The most consequential transitions of this stretch were the Cambrian Explosion that gave the planet its modern animal body plans, the colonisation of the dry continental surface by plants, and the slow rebuilding of vertebrate biology — first into limbed but water-bound amphibians, then into amniotes whose eggs no longer needed water.

By the close of the Carboniferous, the amniotes are about to split. One branch will become the synapsids, the lineage that eventually produces the mammals. The other will become the sauropsids, the lineage that includes turtles, lizards, snakes, crocodiles, the dinosaurs, and the birds. The next entry follows that split forward — through the great Permian extinction, the Mesozoic dominance of the dinosaurs, the long shadow under which the early mammals lived for a hundred and fifty million years, and the asteroid-driven catastrophe at the end of the Cretaceous that opened the world to mammalian radiation.

The vertebrates have made it onto land. The question of which kind of vertebrate would eventually inherit it is the subject of what comes next.