Atlas / Biology / The Tree of Life Thread

Field · Emerged 1796 – 1991

Palaeontology

What can be reconstructed about vanished life, and about the history of the Earth, from the fragments that happen to have survived?

5 chapters7 min read6 turning points 2 open problems

Branched from: Systematics
Branched into: Macroevolution
Figures: Georges Cuvier, Mary Anning, William Smith, Richard Owen, Charles Walcott, Harry Whittington, Stephen Jay Gould, Simon Conway Morris, Derek Briggs, Walter Alvarez, Luis Alvarez

In brief

Fossils were known for as long as people have dug, and for most of that time they were explained as sports of nature, failed creations or the debris of Noah's flood. Palaeontology began when Georges Cuvier showed in 1796 that the fossil elephants of Siberia and Ohio belong to species that no longer exist anywhere. Extinction became a fact, and the Earth acquired a history with chapters.

The subject grew two capabilities that nothing else in biology has. It can see time directly: strata lie in order, and the fossils in them succeed each other in the same order everywhere, which lets rocks on different continents be correlated and, once radioactivity supplied a clock, dated absolutely. And it records events that no living organism witnessed — five mass extinctions, the sudden appearance of animal body plans, and the arrival of an asteroid whose trace is a centimetre of iridium-rich clay laid down worldwide in a single season.

Key ideas

ExtinctionEnters 1796 – 1812: Species can cease to exist. Cuvier established this by comparative anatomy, against the prevailing view that no link in the chain of being could be lost, and it is the premise on which every later reading of the fossil record rests.
Faunal successionEnters 1799 – 1815: Each layer of rock contains a characteristic assemblage of fossils, and the assemblages follow one another in the same sequence wherever they are found. This makes fossils a clock: a stratum anywhere can be placed in the sequence by what it contains.
Relative versus absolute datingEnters 1799 – 1815: Faunal succession gives order without duration — this is older than that, by an unknown amount. Radioactive decay supplies the duration, and the two together turn the column of strata into a calendar.
TaphonomyEnters 1909 – 1989: Everything that happens to a body between death and discovery: decay, scavenging, burial, compaction, dissolution, exposure. The fossil record is the output of this filter, not a sample of the living world, and reading it means modelling the filter.
LagerstätteEnters 1909 – 1989: A deposit of exceptional preservation, where soft tissue survived. These are rare, local and wildly informative: most of what is known about the Cambrian animals comes from a handful of them.

Draws on other domains

↙ Physics
Radiometric Dating
Putting dates on the fossil record

Chapter I

Extinction Becomes a Fact

In 1796 a young anatomist at the Paris museum gave a lecture on elephants. Georges Cuvier compared the jaws and teeth of the Indian and African species — different enough, he argued, to be distinct — and then set beside them the mammoth skulls dug from the Siberian permafrost and the enormous molars sent from a salt lick in Ohio. Neither matched either living elephant. Animals of that size could not be overlooked in some unexplored country. They were gone.

This was not a small claim. It cut against the idea of a complete and permanent chain of being, and it meant the Earth had a history in which whole populations of creatures ended. Cuvier spent the next sixteen years proving it over and over, species by species, in the Recherches sur les ossemens fossiles. He also noticed that the faunas in the Paris basin changed abruptly between strata, and concluded that the history was punctuated by catastrophes — a view his successors rejected and the evidence later partly restored.

The supply of fossils came largely from people outside the lecture halls. Mary Anning sold specimens from the Jurassic cliffs at Lyme Regis to support her family, and in the process produced the first described ichthyosaur, the first complete plesiosaur and the first British pterosaur. She worked out that the coprolites in the blue lias were fossil faeces, and could therefore say what the animals had eaten. Her finds were named and published by Fellows of a society that would not admit her.

Chapter II

Reading Time from Rock

William Smith surveyed canals, which meant walking a great many fresh cuttings. He noticed that each bed carried a distinctive set of shells, and — the crucial part — that the beds always appeared in the same order, whether they were thick or thin, chalk or clay, in Somerset or in Yorkshire. Fossils were therefore a label for position in time. A stratum could be identified anywhere in the country by what was in it, and strata on different continents could be matched.

Faunal succession gives order but no duration. Smith could say that the lias is older than the chalk; he could not say by how much, and estimates of the age of the Earth through the nineteenth century ranged over two orders of magnitude. The duration came from physics. Once radioactive decay was understood, minerals in volcanic ash beds could be dated absolutely, and the relative column became a calendar — a joining of two sciences described under radiometric dating. The Cretaceous–Palaeogene boundary is now placed at 66.0 million years ago with an uncertainty of a few tens of thousands of years, which is precise enough to argue about what happened first.

Chapter III

Monsters, Museums and the Cambrian

In 1842 Richard Owen noticed that three poorly known British fossil reptiles shared a fused sacrum and limbs held beneath the body rather than sprawled, and gave the group a name: Dinosauria. The name did more work than the anatomy. Within twelve years full-sized concrete reconstructions stood in the grounds at Crystal Palace, and palaeontology acquired a public that no other science had.

The deeper surprise came later and higher up. In 1909 Charles Walcott found a Cambrian shale on a ridge in British Columbia in which soft parts were preserved as carbon films — guts, eyes, gills, whole animals that had no hard tissue to leave behind. He placed the fossils in known groups. Sixty years on, Harry Whittington and his students took the specimens apart again and found creatures that resisted placement: Opabinia with five eyes and a grasping proboscis, Anomalocaris, Hallucigenia. Stephen Jay Gould made the case in Wonderful Life that the Cambrian had run more experiments in body plan than have existed since, and that the survivors were chosen by luck as much as by fitness. Simon Conway Morris, who had done much of the redescription, argued the opposite: that the oddities are early branches of lineages that still exist, and that convergence makes life's outcomes largely predictable. The disagreement is live and is as much about how to measure difference as about the animals.

Chapter IV

A Closer Look: Weighing an Asteroid from a Centimetre of Clay

At Gubbio in Italy the Cretaceous limestones end, a centimetre of clay intervenes, and the Palaeogene limestones begin. Walter Alvarez wanted to know how long the clay took to accumulate, and his father Luis Alvarez suggested a clock: iridium. It is almost absent from the Earth's crust, which was stripped of siderophile elements when the core formed, but it is present in meteorites at around 500 parts per billion, and it rains down steadily as micrometeorite dust. A slowly deposited clay should hold more of it than a fast one.

The measurement, by Frank Asaro and Helen Michel, found about 9 parts per billion against a background near 0.3 — thirty times too much, and far more than slow deposition could explain. The same spike appeared at Stevns Klint in Denmark and in New Zealand. Suppose instead that it arrived all at once, as a vaporised impactor spread through the atmosphere and settled worldwide. How big was it?

Take the layer as 1 cm thick with a density of about 2 g/cm³, so it holds 2 grams of material per square centimetre of ground. At 9 parts per billion, each square centimetre carries

2 \text{ g} \times 9 \times 10^{-9} = 1.8 \times 10^{-8} \text{ g of iridium}.

The Earth's surface is $4\pi R^2$ with $R = 6.371 \times 10^{8}$ cm, or $5.1 \times 10^{18}$ cm². The total excess iridium is therefore

1.8 \times 10^{-8} \times 5.1 \times 10^{18} \approx 9.2 \times 10^{10} \text{ g},

about ninety thousand tonnes of one of the rarest metals on Earth, deposited in a single layer. If it came from a chondritic body containing 500 ppb iridium, that body massed

\frac{9.2 \times 10^{10}}{5 \times 10^{-7}} \approx 1.8 \times 10^{17} \text{ g} = 1.8 \times 10^{14} \text{ kg}.

At a density of 2.2 g/cm³ this is a volume of $8.4 \times 10^{16}$ cm³, and a sphere of that volume has diameter

d = \left(\frac{6V}{\pi}\right)^{1/3} \approx 5.4 \times 10^{5} \text{ cm} \approx 5 \text{ km}.

The Alvarez paper arrived at about 10 km, allowing for iridium spread beyond the boundary clay itself and for material lost to space. Chicxulub, the 180-km crater found under the Yucatán in 1991, implies the same scale. Either way the kinetic energy is the part that matters. At a typical encounter speed of 20 km/s, a body of $1.8 \times 10^{14}$ kg carries

E = \tfrac{1}{2}mv^{2} = \tfrac{1}{2}\left(1.8 \times 10^{14}\right)\left(2 \times 10^{4}\right)^{2} \approx 3.6 \times 10^{22} \text{ J},

which is about nine million megatons of TNT, some six hundred million times the Hiroshima bomb, delivered in seconds. The argument is worth following because of what it is made of: a thickness, a density, a concentration and a sphere. No fossils are involved at all. A quantity measured in parts per billion in one Italian roadcut, multiplied by the area of a planet, weighs an object that no one saw.

Chapter V

What the Record Can and Cannot Say

Every fossil is a survivor of a filter. An animal must die in the right place, be buried fast, escape dissolution and compaction, sit through tens of millions of years of burial, and then be exhumed where somebody is looking. Soft-bodied life passes this filter almost never, which is why the Burgess Shale and a handful of other deposits carry so much of the weight. Counts of species through time partly track the amount of rock of each age that happens to be exposed, and the last fossil of a declining species always predates its real end, so a sudden extinction is recorded as a gradual one.

Modelling that filter well enough to invert it is the central methodological problem of the subject, and it is why macroevolution — the study of what the record means over hundreds of millions of years — had to wait for databases and statistics. The other route past the filter is to stop reading morphology and start reading molecules, which is where phylogenetics comes in: living organisms carry information about their dead ancestors, and a tree built from them can be calibrated against the few dates that stone supplies.

Turning points6

1796 – 1812CONSENSUS-OVERTURNED
Cuvier proves that species die out
In a lecture to the Institut National in Paris, Georges Cuvier compares the jaws and molars of Indian and African elephants with the fossil mammoth of Siberia and the "Ohio animal" later named mastodon, and shows that the fossils are neither of the living species. Since animals that large could not be hiding, they must be gone. His Recherches sur les ossemens fossiles of 1812 extends the argument to dozens of species and divides the strata into successive faunas separated by catastrophes.
Georges Cuvier
› Sources (2)
- Cuvier, G. (1796). Mémoire sur les espèces d'éléphans tant vivantes que fossiles. Magasin Encyclopédique 2: 440–445.
- Rudwick, M. J. S. (1997). Georges Cuvier, Fossil Bones, and Geological Catastrophes. University of Chicago Press.
1811 – 1828DISCOVERY
The reptiles of the Lyme Regis cliffs
Mary Anning, who collected fossils commercially from the crumbling Jurassic cliffs of Dorset, finds the first ichthyosaur skeleton to be properly described, in 1811–12 with her brother Joseph; the first complete plesiosaur, in 1823; and the first British pterosaur, in 1828. She also recognised that bezoar stones were fossil faeces, which let her reconstruct what the animals ate. The specimens were described and named by men of the Geological Society, from which, as a woman and a tradesperson, she was excluded.
Mary Anning
› Sources (2)
- Torrens, H. (1995). Mary Anning (1799–1847) of Lyme; 'the greatest fossilist the world ever knew'. British Journal for the History of Science 28(3): 257–284.
- Emling, S. (2009). The Fossil Hunter. Palgrave Macmillan.
1799 – 1815DISCOVERY
Strata identified by their fossils
William Smith, a canal surveyor walking the cuttings of southern England, notices that each bed of rock carries its own set of fossils and that the beds always occur in the same order, whatever their thickness or lithology. In 1815 he publishes A Delineation of the Strata of England and Wales, a hand-coloured geological map on eight feet of paper. It made the fossil record a correlation tool and the basis of economic geology.
William Smith
› Sources (2)
- Smith, W. (1815). A Delineation of the Strata of England and Wales, with part of Scotland. John Cary, London.
- Winchester, S. (2001). The Map That Changed the World. HarperCollins.
1842SYNTHESIS
Dinosauria
Richard Owen, reviewing British fossil reptiles, notices that Megalosaurus, Iguanodon and Hylaeosaurus share features — fused sacral vertebrae, limbs held under the body — that set them apart from lizards, and erects the group Dinosauria for them. The name gave a scattered set of bones an identity, and the public exhibitions that followed made palaeontology the first science with a mass audience.
Richard Owen
› Sources (2)
- Owen, R. (1842). Report on British fossil reptiles, part II. Report of the British Association for the Advancement of Science 11: 60–204.
- Torrens, H. (1992). When did the dinosaur get its name? New Scientist 134: 40–44.
1909 – 1989DISCOVERYContested
The Burgess Shale and the problem of disparity
Charles Walcott, high on a ridge in British Columbia, finds a Cambrian deposit in which soft tissue is preserved as carbon films: eyes, guts, gills, animals with no hard parts at all. He assigned the fossils to known groups. From 1971 Harry Whittington, Derek Briggs and Simon Conway Morris redescribed them, and some — Opabinia with five eyes, Hallucigenia walking on spines — fitted nowhere. Stephen Jay Gould's Wonderful Life (1989) argued that the Cambrian held more fundamentally different body plans than exist today, and that which ones survived was largely luck.
What is disputed
Whether early Cambrian disparity really exceeded modern disparity is disputed. Conway Morris, whose redescriptions Gould relied on, argued that most of the oddities are stem members of surviving groups rather than vanished phyla, and that convergence makes evolutionary outcomes far more predictable than Gould's "replay the tape" image allows. Quantitative studies of morphospace have supported each side at different taxonomic scales.
Charles Walcott, Harry Whittington, Stephen Jay Gould, Simon Conway Morris, Derek Briggs
› Sources (3)
- Gould, S. J. (1989). Wonderful Life: The Burgess Shale and the Nature of History. W. W. Norton.
- Conway Morris, S. (1998). The Crucible of Creation. Oxford University Press.
- Briggs, D. E. G., Erwin, D. H. & Collier, F. J. (1994). The Fossils of the Burgess Shale. Smithsonian Institution Press.
1980 – 1991DISCOVERYContested
An asteroid at the end of the Cretaceous
Walter Alvarez, dating the thin clay at the Cretaceous–Palaeogene boundary near Gubbio, asks his father Luis Alvarez how long it took to form. Measuring iridium, which is rare in the crust and common in meteorites, Frank Asaro and Helen Michel find a spike thirty times background — and the same spike turns up in Denmark, New Zealand and everywhere else the boundary is exposed. The 1980 paper proposes an impact by a body about 10 km across. The crater, 180 km wide and buried under the Yucatán, was identified in 1991.
What is disputed
That a large impact occurred is no longer disputed. Whether it alone caused the extinction is. The Deccan Traps in India erupted hundreds of thousands of cubic kilometres of lava across the same interval, and a substantial minority of palaeontologists argue that volcanic climate disruption had already destabilised ecosystems, with the impact as the final blow. Dating of both events has narrowed but not closed the question.
Walter Alvarez, Luis Alvarez
› Sources (3)
- Alvarez, L. W., Alvarez, W., Asaro, F. & Michel, H. V. (1980). Extraterrestrial cause for the Cretaceous–Tertiary extinction. Science 208: 1095–1108.
- Hildebrand, A. R. et al. (1991). Chicxulub crater: a possible Cretaceous/Tertiary boundary impact crater on the Yucatán Peninsula, Mexico. Geology 19: 867–871.
- Schoene, B. et al. (2019). U-Pb constraints on pulsed eruption of the Deccan Traps across the end-Cretaceous mass extinction. Science 363: 862–866.

Applications

Where it is used

Energy and resources
Dating a well by its microfossils
Biostratigraphy is routine industrial practice. Drilling returns chips of rock, and the foraminifera, pollen and nannofossils in them identify which formation the bit is in and how deep the target lies. The method is Smith's, unchanged in principle since 1815, applied to organisms a tenth of a millimetre across.
› Sources (1)
- Jones, R. W. (2006). Applied Palaeontology. Cambridge University Press.
Climate science↗ Physics · Geophysical Fluid Dynamics
Past climates as a test of models
Oxygen isotope ratios in foraminiferal shells record ocean temperature, and leaf shape, pollen assemblages and tree rings record conditions on land. The Palaeocene–Eocene Thermal Maximum, a carbon release 56 million years ago, is used as a natural experiment against which climate models are checked at warming far beyond the instrumental record.
› Sources (1)
- Zachos, J. C., Dickens, G. R. & Zeebe, R. E. (2008). An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature 451: 279–283.

Open problems

Where the map runs out

Open

Correcting the record for its own gaps

Open as of 2026; competing correction methods still give different diversity curves.

Raw counts of fossil species through time track the amount of rock available, the number of palaeontologists who have worked on an interval, and the sampling effort, as much as they track diversity. Every claimed long-term pattern — the rise of marine diversity since the Permian, the shape of recoveries after extinctions — depends on how the counts are corrected, and the main correction methods disagree.

Why it is hard: The bias is not noise but a structured filter that varies with time, place, body type and environment. Soft-bodied animals are missing almost entirely; shallow marine shells are over-represented; the last appearance of a rare species is always earlier than its true extinction, which smears sudden events into gradual ones.
What resolving it unlocks: Whether diversity has risen steadily, saturated, or cycled is the central empirical question about the history of life, and it cannot be settled until the filter is modelled well enough to be inverted.

› Sources (2)

Alroy, J. et al. (2008). Phanerozoic trends in the global diversity of marine invertebrates. Science 321: 97–100.
Signor, P. W. & Lipps, J. H. (1982). Sampling bias, gradual extinction patterns, and catastrophes in the fossil record. Geological Society of America Special Paper 190: 291–296.