A When the Dutch navigator Jacob Roggeveen sighted a small, almost treeless island in the south-eastern Pacific on Easter Sunday 1722, he named it after the day and recorded little more than his astonishment at the stone figures lining its shores. The island, which its own people call Rapa Nui, covers barely 164 square kilometres and lies some 3,500 kilometres from the coast of South America, making it one of the most remote inhabited places on the planet. In the three centuries since, outsiders have tried to explain how so small and isolated a community produced nearly 900 monumental statues, known as moai, and what became of the society that carved them. The answers proposed have often revealed as much about the assumptions of those who offered them as about the island itself.

B Early European visitors were inclined to doubt that the islanders could have been responsible at all. When the British navigator James Cook called in 1774, he found a population that seemed too small and too poorly equipped to have moved such enormous figures, and several of the statues had already been toppled. Speculation about a vanished race of master builders followed. It was only with the expedition of the British archaeologist Katherine Routledge in 1914–15 that a systematic attempt was made to survey the moai and to record what elderly islanders remembered of their past. Her notebooks, which preserved genealogies and place names that might otherwise have been forgotten, remain a valuable source more than a century later.

C In the mid-twentieth century the Norwegian adventurer Thor Heyerdahl offered a more dramatic account. Convinced that the island had been settled from South America, he sailed a balsa-wood raft, the Kon-Tiki, from Peru into Polynesia in 1947 to show that such a voyage was possible, and later led excavations on Rapa Nui itself. His theory won a wide popular following but little support among specialists, for whom the islanders' language, crops and tools pointed unmistakably westwards, to the rest of Polynesia. Modern genetics has largely confirmed the specialists' view, although it has also complicated it: a 2024 study of ancient genomes found that the Rapanui carry a modest share of Native American ancestry, introduced through contact that probably took place several centuries before any European ship appeared. Heyerdahl, in other words, was wrong about who the first settlers were, but not entirely wrong that the two worlds had met.

D The date of settlement has been revised as well. Radiocarbon results obtained from the 1950s onwards suggested that people had arrived as early as the fourth or fifth century AD. More rigorous dating of charcoal from undisturbed layers, published by the archaeologists Terry Hunt and Carl Lipo in 2006, pointed instead to a much later arrival, around 1200. The earlier figures, they argued, had relied on samples that were poorly located or contaminated. If the later date is correct, the islanders created their extraordinary monuments within a few hundred years of landing, a far shorter span than had once been supposed.

E Most controversial of all has been the question of decline. In his best-selling book Collapse (2005), the geographer Jared Diamond presented Rapa Nui as a cautionary tale: the islanders, he argued, cleared their palm forests, partly to obtain the timber and rope needed for transporting statues, until soil erosion, food shortages and warfare tore their society apart. Pollen preserved in lake sediments does show that the island was once heavily wooded and that the forest had largely disappeared by the time Europeans arrived. What is disputed is the cause and the consequences. Hunt and Lipo pointed out that the shells of the native palm's nuts, recovered from caves and excavations, almost always bear the gnaw marks of the Polynesian rat, which travelled with the first settlers and, lacking predators, may have multiplied rapidly enough to prevent the trees from regenerating.

F Evidence of a catastrophic fall in population before European contact has also proved elusive. The same 2024 genetic study found no sign of the sharp narrowing of genetic diversity that a seventeenth-century collapse would be expected to leave behind. Another study published that year used satellite imagery and machine learning to map the island's 'rock gardens', plots in which stones were scattered over the earth to keep it moist and fertile. The gardens turned out to cover far less ground than earlier estimates had assumed, implying that, even with fish and other foods added to the diet, the island could have supported no more than about 4,000 people, rather than the 15,000 or more suggested in some earlier work. A modest population that remained broadly stable for centuries is difficult to reconcile with the idea of self-inflicted ruin.

G There is no doubt, however, that the island's society suffered devastating losses once outsiders began to arrive. In 1862 and 1863, slave raiders from Peru carried off around 1,500 people to labour on the mainland. Most died within months, and the handful of survivors eventually sent home brought smallpox with them. Combined with other introduced diseases and emigration, these events left only 111 islanders by 1877. For many scholars, this is where the true collapse lies, and they argue that the popular image of a people who wrecked their own environment has drawn attention away from it. The debate is far from settled, but it has gradually shifted the focus from what the islanders did to their island to what was later done to them.

A To most visitors, the stalagmites rising from the floors of limestone caves are simply a spectacle, the patient sculpture of dripping water. For much of the twentieth century scientists regarded them in much the same way. Over the past three decades, however, these formations, together with the stalactites that hang above them and related mineral deposits known collectively as speleothems, have become one of the most valuable sources of information about the Earth's past climate. Unlike ice cores, which can only be drilled where glaciers and ice sheets survive, they occur across a vast range of latitudes and altitudes, including the tropics and the interiors of continents.

B The process that produces them begins above ground. As rainwater soaks through the soil, it absorbs carbon dioxide released by plant roots and decaying organic matter. The result is a weak solution of carbonic acid, which dissolves calcium carbonate from the limestone as it seeps downwards through cracks and joints in the rock. When a drop finally emerges from the cave ceiling, it meets air containing far less carbon dioxide than the soil it passed through. The water therefore loses some of its dissolved gas, can no longer hold its entire mineral load, and deposits a microscopic film of calcite. Where drops fall to the floor, the same exchange builds a stalagmite upwards. Growth is extremely slow, typically in the order of a tenth of a millimetre a year, so a specimen a metre tall may represent ten thousand years or more of accumulation.

C What turned speleothems into scientific instruments was the ability to date them precisely. Drip water carries tiny quantities of dissolved uranium, which become locked into the growing calcite. Thorium, by contrast, is virtually insoluble in water, so newly formed calcite contains almost none. Over time, the trapped uranium decays into thorium at a known and constant rate, and by measuring the ratio between the two, researchers can calculate how long ago each layer formed. Refinements in mass spectrometry, introduced in the late 1980s by the geochemist R. Lawrence Edwards and his colleagues, sharply reduced the size of sample required and improved accuracy, allowing layers up to about half a million years old to be dated to within a few decades or centuries.

D Dating alone would be of limited use without a record of climate to attach it to, and that record comes largely from oxygen. Water molecules contain either the common, lighter form of oxygen or a rarer, heavier isotope, and the proportion of the two in the calcite mirrors that of the drip water, and hence of the rain that fell above the cave. In monsoon regions, periods of intense rainfall tend to produce water relatively poor in the heavier isotope. In 2001 a team led by Yongjin Wang of Nanjing Normal University, working with Edwards and Hai Cheng, published an analysis of stalagmites from Hulu Cave in eastern China. Their record of the East Asian monsoon showed abrupt shifts that closely matched the temperature swings preserved in Greenland ice, indicating that sudden changes around the North Atlantic had been felt many thousands of kilometres away. Later work at Hulu and at Sanbao Cave in central China extended the record back some 640,000 years.

E Precisely dated cave records have also been drawn into debates about human history. In 2012 a team led by the archaeologist Douglas Kennett analysed a stalagmite from Yok Balum Cave in Belize and reported that a series of severe droughts coincided with the political disintegration of Classic Maya cities between roughly AD 800 and 1000. A stalagmite from Mawmluh Cave in the Indian state of Meghalaya, meanwhile, preserves evidence of an abrupt drying around 4,200 years ago, an episode that some researchers have linked to the decline of early civilisations in Egypt, Mesopotamia and the Indus Valley. In 2018 the International Commission on Stratigraphy selected that same stalagmite as the reference point for the beginning of the Meghalayan Age, the most recent division of the present geological epoch.

F Yet the record in the stone is not always easy to read. The isotopic composition of rainfall depends not only on how much rain falls but also on where the moisture originated and the route storms took, and a number of climate scientists have argued that the Chinese records partly reflect such changes in moisture source rather than local rainfall alone. Water may also take years or even decades to percolate through the rock, blurring short-lived events, and if a cave is poorly ventilated or a drip falls too rapidly, the calcite may form without faithfully recording the chemistry of the water. Researchers therefore monitor drip rates, temperature and air composition in the caves they study, and place greater confidence in a result when two or more stalagmites from the same cave yield matching records, since a common pattern is unlikely to arise from the peculiarities of a single drip.

G Such sampling raises an ethical problem of its own. A stalagmite that took tens of thousands of years to grow cannot be replaced, and slicing one in half to analyse its layers destroys a feature that cavers and tourists may value as much as scientists do. Many researchers now follow codes of practice that favour formations which have already fallen or broken, the extraction of narrow cores rather than whole specimens, and close cooperation with landowners and cave managers. The aim is to ensure that an archive assembled over millennia is not destroyed in the act of reading it.

A Few industries have embraced the vocabulary of sustainability as eagerly as fashion. Retailers invite customers to drop unwanted garments into collection bins in exchange for discount vouchers, advertise ranges made from 'recycled' fibres, and speak of a 'circular' future in which yesterday's clothes are remade into tomorrow's. The appeal of the idea is obvious: if old garments can simply be turned into new ones, then the sheer volume of clothing the world buys need not be a problem at all. Yet the evidence suggests that recycling, at least as currently practised, offers far less than this rhetoric implies, and that relying on it risks postponing the harder choices the industry faces.

B The scale of the challenge is difficult to overstate. According to a 2017 report by the Ellen MacArthur Foundation, a charity that promotes the circular economy, global clothing production roughly doubled between 2000 and 2014, while the average number of times a garment was worn before being discarded fell by more than a third over a similar period. The same report estimated that less than one per cent of the material used to make clothing is recycled into new clothing. A review published in 2020 by the Finnish design researcher Kirsi Niinimäki and her colleagues put the industry's annual output of textile waste at more than 92 million tonnes and its yearly water consumption at around 79 trillion litres. Whatever the precise figures, the direction of travel is unmistakable.

C Part of the difficulty is technical. Most garments are made not of a single material but of blends, such as cotton mixed with polyester or elastane, and separating these fibres is costly and, with current methods, often impossible. Mechanical recycling, which shreds fabric back into fibre, shortens cotton fibres so that they usually have to be combined with virgin material to produce yarn of acceptable quality; each cycle, in effect, degrades the resource. Chemical processes that dissolve fabrics and rebuild their fibres promise better results, but they have struggled to compete on price. Renewcell, a Swedish company widely celebrated for turning discarded cotton into a pulp that could be spun into new fibre, filed for bankruptcy in early 2024, having failed to secure enough customers. The technology worked; the market for it did not.

D Much of what is labelled recycled, moreover, has never been clothing at all. Textile Exchange, an industry body that tracks fibre use, has reported that the overwhelming majority of recycled polyester is made from plastic bottles. Turning bottles into fleeces may divert some plastic from landfill, but it does nothing to reduce the quantity of garments being thrown away, and it arguably draws bottles out of a system that could have made them into bottles again. To present such products as evidence of a circular fashion industry is, at best, generous.

E Nor do collection schemes necessarily achieve what their promoters claim. Only a fraction of donated clothing is resold in the country where it was given; much of the remainder is exported, often to markets in Africa, Asia and Eastern Europe. At Kantamanto in Accra, Ghana, one of the world's largest second-hand clothing markets, the OR Foundation, a non-profit organisation that works with local traders, has estimated that around 40 per cent of the garments in each imported bale are of such poor quality that they end up as waste, much of it dumped or burned. Meanwhile, critics argue that offering vouchers in return for old clothes may encourage shoppers to buy more, easing any guilt they feel about consumption rather than curbing it. That suspicion seems to me entirely reasonable: a scheme designed to bring customers back into a shop is, after all, a marketing device as much as an environmental one.

F It is tempting to conclude that the solution lies with consumers, who could simply buy fewer, better things. The design researcher Kate Fletcher has argued persuasively that how garments are used, mended and kept matters as much as how they are produced, and her work has helped to inspire a lively movement around repair and 'slow fashion'. But placing the burden chiefly on individual shoppers is, I believe, a mistake. Consumer choices are shaped by prices, advertising and the sheer availability of cheap clothing, and appeals to personal virtue have so far done little to change behaviour across society as a whole.

G A more promising approach is to change the incentives facing producers. France has operated an extended producer responsibility scheme for textiles for more than fifteen years, requiring brands to contribute to the cost of collecting and processing what they sell, and its 2020 anti-waste law went further by prohibiting the destruction of unsold clothing and other non-food goods. The European Union has required member states to collect textiles separately from 2025 and has agreed that producer-funded schemes of this kind should become compulsory across the bloc. Such measures are not glamorous, but by attaching a cost to every garment placed on the market, they give manufacturers a reason to make fewer, more durable and more easily recycled products.

H None of this means that recycling should be abandoned. Better sorting technology, designs that avoid inseparable blends and investment in fibre-to-fibre processes will all be needed. But recycling is a way of managing waste, not of preventing it, and an industry that produces ever-larger volumes of clothing cannot recycle its way to sustainability. A circular economy, if the phrase is to mean anything, must begin with the question of how much is made in the first place.