How old is the Earth?

A night in 4004 BC

According to James Ussher, Archbishop of Armagh in Ireland, the world began in the evening before Sunday, 23 October 4004 BC. His great chronology, Annales veteris testamenti (“Annals of the Old Testament”), appeared in Latin in London in 1650. He had worked out the date with a scholar’s care. He anchored his count with a fixed point, the death of the Babylonian king Nebuchadnezzar in 562 BC, known from Babylonian king-lists and Ptolemy’s canon of kings. He counted back through the generations of Genesis, using the ages given in the Hebrew text. And he used astronomical tables to find the first Sunday after the autumn equinox in that year. That Sunday was 23 October, and Ussher concluded that creation began “at the start of the evening preceding” it. (He gave no clock time. His own arithmetic, 4,003 years, 70 days and 6 hours from that evening to midnight at the start of AD 1, implies about 6 p.m. on the Saturday, sunset. The “9 a.m.” of popular retellings belongs to a rival scholar, John Lightfoot, and refers to the making of man.)

Ussher is often mocked today, and he should not be. His method was serious, and his answer was not an oddity. Rival scholars, among them Joseph Scaliger, Lightfoot, and Isaac Newton, proposed creation dates within about a century of his, roughly 3928 to 3998 BC depending on the list; a later count found more than two hundred different creation dates proposed from the Bible alone. His dates did not enter the printed English Bible until 1701, long after he had died, when they appeared in the margins of a folio edition. They were never part of the text.

The idea that the world is only a few thousand years old came from one way of reading Genesis. The early Christian chronologists had made the sums in various ways: Julius Africanus, about the year 221, counted roughly 5,500 years from creation to Christ; Eusebius, a century later, counted 2,242 years from Adam to the Flood in the Greek translation of the Bible, where the Hebrew text gives fewer; Martin Luther’s chronology of 1541 put creation at about 3961 BC; and the German astronomer Johannes Kepler argued that Christ had been born a few years before the year 1. Ussher took Christ’s birth to be 4 BC, because Herod died earlier than AD 1.

Other traditions had other ideas. Aristotle argued that the heavens have no beginning and no end: the world has always been. Augustine, in the fifth century, warned Christians against talking nonsense about nature in front of unbelievers; Galileo quoted him in 1615. In China, the philosopher Shao Yong (1011–1077) described a great cycle of 129,600 years, and in Hindu tradition a “day of Brahma” is 4.32 billion years. These are cycles of world history, however, built from numerology, and not measured ages of the earth. That one of them comes close to the modern figure is a coincidence of order of magnitude only. Newton, who devoted much of his time to chronology, was working on human history: his Chronology of Ancient Kingdoms Amended (1728) tried to shorten the dates of Greek history. His surviving papers are dominated by theology, prophecy, chronology, and alchemy, and the economist John Maynard Keynes called him “the last of the magicians.”

Shells on the mountain

Meanwhile, people kept finding seashells on mountains. In China the observation is old and the interpretations were bold. In AD 771 the calligrapher and official Yan Zhenqing, writing about the Magu mountain in Jiangxi, noted that high up among the rocks one still finds the shells of clams and oysters, and that some took this as a sign that the mulberry fields had once been sea. In the 1070s and 1080s the polymath Shen Kuo, on a mission through Hebei, saw shells and egg-shaped pebbles in horizontal bands in the cliffs of the Taihang mountains. “This was once the seashore,” he wrote; the coast was now nearly a thousand li away (a Song-dynasty li was about half a kilometer), and he explained the plain between as silt from rivers. He also wrote that the peaks at Yandang were left standing after floods washed the soil away, and that a grove of petrified bamboo, found when a riverbank collapsed near Yanzhou (today’s Yan’an), suggested that the ground there had once been lower and damper, as bamboo likes. In the twelfth century the philosopher Zhu Xi, who also noticed shells on high mountains, concluded that what was once low becomes high and what was soft becomes hard. These were observations of land and sea changing places, not measures of age. But they came centuries before Hutton.

In Europe the turning point was a shark. In 1666 the head of a great white shark was brought to the court of the Medici in Florence. The Danish anatomist Niels Steensen (Steno) dissected it and saw that its teeth matched the “tongue stones” that people found in rocks. They were the teeth of ancient sharks. In 1669 he published the first principles of stratigraphy: layers of rock are laid down one on top of another, flat, so that the lower are older. Steno, who converted to Catholicism, became a priest and later a bishop. Robert Hooke in London argued from fossils and earthquakes that parts of the sea floor had been raised into land. Others built theories of the whole earth. René Descartes, in 1644, imagined that the earth had begun as a sun-like fire that cooled and formed a crust; Gottfried Leibniz wrote something similar around 1691–93, though his book appeared only in 1749. Whether the rocks were laid down in water or came from fire became a long argument, between the “Neptunists,” who held that everything settled from a primeval ocean, and the “Plutonists,” who held that heat made the igneous rocks. It ran into the early 1800s.

The most famous early attempt to put a number on a cooling earth was the French naturalist Buffon’s. He made small iron and mineral spheres at his forges at Montbard, heated them, and timed how long they took to cool. In Les époques de la nature (1778) he scaled the result up to the size of the earth: about 75,000 years (the precise figure usually given is 74,832) since it began to cool, and in private notes far more. He knew how it would look. A generation earlier, in 1751, the theology faculty of the Sorbonne had objected to the first volume of his Histoire naturelle, and he had answered that his theory was a purely philosophical supposition. In the Époques he put the matter bluntly: blind respect for old traditions, he wrote, would be superstition.

The abyss of time

Buffon had a number. James Hutton, a Scottish gentleman, had a different idea: that the earth’s age should not be a number at all. Hutton took his medical degree at Leiden in 1749 and never practiced; he farmed the family land in Berwickshire and owned a share in a factory that made sal ammoniac. In 1785 he read a paper on the Theory of the Earth to the Royal Society of Edinburgh (it was printed in 1788). Rocks, he argued, are made of the worn-down remains of older rocks, laid down on the sea floor, hardened by heat and pressure, lifted up into mountains, and worn down again, in a cycle with no end that we can see. In Glen Tilt, in the Highlands, he saw veins of granite running through older rock, which showed that the granite had once been molten. The paper ends with the sentence that made him famous: in the earth “we find no vestige of a beginning,—no prospect of an end.”

In 1788 Hutton went to Siccar Point on the coast east of Edinburgh, in a boat, with his friend the mathematician John Playfair and Sir James Hall. There the cliff shows tilted gray beds standing nearly on end, topped by gently sloping red sandstone: an unconformity, a surface where a huge span of time is missing. The lower rocks were laid down as flat sea-floor mud, hardened, tilted, raised, and eroded; then new layers were laid across them. Today we know that the gap is about 65 million years. Playfair wrote, years later, of how the sight affected him: “The mind seemed to grow giddy by looking so far into the abyss of time.”

The picture filled in quickly. The French anatomist Georges Cuvier showed in 1796 that mammoths and other fossil animals were extinct species, not creatures hiding in unexplored lands; he thought the earth had been through several sudden “revolutions,” the latest only a few thousand years back. The English surveyor William Smith, working in canal-building and mines, found that each layer of rock has its own characteristic fossils, so that fossils can identify layers over great distances; in 1815 he published the first detailed geological map of a whole country, England and Wales. And Charles Lyell’s Principles of Geology (1830–33) argued that the past should be explained by causes still at work today, slow erosion and the slow raising of land, given enough time. Lyell avoided numbers, saying only that the time must be “indefinitely” long.

Charles Darwin carried Lyell’s first volume on the Beagle, and he needed time as much as anyone. In the first edition of On the Origin of Species (1859), he estimated how long the sea and rain would have needed to wear away the great valley of the Weald in southern England, between the North Downs and South Downs. Assuming the rock eroded at about an inch a century, he got “306,662,400 years; or say three hundred million years.” That figure was for the wearing-away of one valley since the chalk was laid down, not for the age of the earth, and it was a rough guess; he deleted it from the third edition after a critical review. But it showed what evolution by natural selection seemed to demand: not thousands of years but hundreds of millions.

A clock that ran fast

The most formidable objection to a very old earth came from physics. William Thomson, later Lord Kelvin, was the best-known physicist in Britain. In 1862 he asked how long the earth could have been cooling. Take the earth as a solid ball of rock that began molten and has lost its heat by conduction, as the French mathematician Joseph Fourier had worked out for a cooling body. Then the age depends on three numbers: the starting temperature, how fast the temperature rises as one goes down (the gradient near the surface), and how well rock conducts heat. Kelvin took a starting temperature of 7,000 degrees Fahrenheit above the surface (about 3,900 °C), the measured gradient of one degree Fahrenheit for every 50 feet of depth, and a conductivity from tests on Edinburgh rocks. His formula gave about 98 million years, with limits of 20 and 400 million. In 1897 he cut it to between 20 and 40 million.

Darwin was troubled. In a letter to Alfred Russel Wallace in April 1869 he wrote that Thomson’s views on the recent age of the world had been “one of my sorest troubles.” Thomas Henry Huxley answered in public that mathematics is like a mill: what comes out depends on what you put in. In 1895 the engineer John Perry pointed out that if the earth’s interior flows, it carries heat to the surface far faster than conduction does, which allows a much older earth. Kelvin did not accept it. And in 1899 the American geologist Thomas Chamberlin warned that Kelvin’s premises were open to doubt: could anyone be sure that “no unrecognized sources of heat” lay inside the earth, or that atoms were not “seats of enormous energies”?

Chamberlin was right, four years before the evidence. Henri Becquerel found radioactivity in 1896 and Marie and Pierre Curie isolated radium in 1898; in 1903 Pierre Curie and Albert Laborde showed that radium gives off enough heat to melt its own weight of ice in under an hour, continuously. Ernest Rutherford, lecturing in 1904 with the aged Kelvin in the audience (by his own later account), argued that radioactive decay supplies heat inside the earth, and that Kelvin’s limit stood only if no new source of heat were found. (The source was radioactive decay, not “nuclear reactions.”) Kelvin never conceded.

Radioactivity did more than rescue Darwin: it supplied a clock. Atoms of uranium decay into lead at a fixed rate, whatever the temperature or pressure, so that in a mineral that has stayed closed since it formed, the ratio of lead to uranium measures its age. In 1907 the American chemist Bertram Boltwood dated twenty-six minerals, and his first results spanned 92 to 570 million years (recalculated later to about 410 million to 2.2 billion). In 1911 the young English geologist Arthur Holmes, who would spend his career refining the clock, got about 1.6 billion years for a mineral from Ceylon, and in 1913 published The Age of the Earth. By 1946 Holmes and, independently, the German physicist Fritz Houtermans used lead isotopes to estimate roughly three billion years for the earth.

The age of the earth, as opposed to its oldest rocks, was fixed by a young geochemist at the California Institute of Technology. Rocks on the earth’s surface are recycled by erosion, melting, and the motion of the crust; the oldest we know is about 4 billion years, and the oldest mineral grains, tiny zircon crystals from Western Australia, about 4.4 billion. The planet itself is older. So Clair Patterson turned to meteorites, leftovers from the same cloud of gas and dust from which the solar system formed, which were never reworked. He chose a piece of the Canyon Diablo meteorite, an iron meteorite from the crater in Arizona, whose sulfide mineral holds almost no uranium and so preserves the lead it was born with. Using a mass spectrometer, with colleagues at Caltech and the Argonne National Laboratory, he compared its lead with the lead in other meteorites and in modern ocean mud. They fell on a single line, as lead from one origin would. The analysis was finished in 1953, and in 1956 Patterson published the result: 4.55 billion years, give or take 70 million. It has barely changed. Today the age of the earth is given as 4.54 billion years, with an uncertainty of about 50 million, and the oldest solid material in the solar system, small mineral inclusions in meteorites, is dated at about 4.567 billion.