Archaeologists are discovering much about ancient peoples by what was on their menus. Jerome Burne reports

When you go round a museum, the last thing you expect is the smell of cooking. Archaeology deals in the currency of objects - bones, tools and carvings - so bleached and stained with age that anything so sensual as an aroma is long gone. The scents and flavours of the past are left to the imagination. But Pennsylvania University's Museum of Archaeology and Anthropology is arranging special dinners where you can smell and eat food similar to that served to the mourners at the funeral of King Midas 2,718 years ago.

That is the same Midas of the legendary golden touch who was a king of Phrigia (now Turkey), and whose huge 230-foot funeral mound was first excavated in 1957. Thanks to some high-tech wizardry involving mass spectrometers, Patrick McGovern has been able to analyse the dusty residue on the cooking pots that were buried with the king to determine what was on the menu at his send-off.

This is one of the more dramatic examples of a molecular revolution that is transforming archaeology. Thanks to increasingly sophisticated hardware, scientists are able to detect the chemical signature of the fats and proteins found in food and reconstruct meals eaten by people thousands of years ago.

ÖйúAƬ

In recent months, these techniques have ended a long-running dispute about the lives of farmers in northwest Scotland and triggered an intense debate about the eating habits of Otzi, the Neolithic man whose 5,300-year-old body was found preserved in an Alpine glacier a few years ago. Scientists with test tubes drawn have been arguing over whether he was a carnivore or a vegan.

The desiccated residue from King Midas's pots had lain untouched on a shelf at Pennsylvania University for about 50 years because, when first uncovered, the tools to analyse them did not exist. McGovern found them and used a spectroscope to investigate the dust taken from the drinking vessels. The molecules that showed up, such as tartaric acid indicating grapes, and calcium oxalate, a by-product of beer, told him he was looking at the remains of a Bronze Age punch, consisting of wine, beer and mead.

ÖйúAƬ

Identifying the more complex compounds in the food remnants needed more sophisticated equipment. High-performance gas and liquid chromatography revealed a pattern of cholesterol, tricylglycerols and fatty acids that gave McGovern the dinner menu from 718BC - barbecued sheep or goat, basted with olive oil, honey and spices, which was then put into a stew with lentils, honey, wine and fennel.

Knowing details of what people ate does not just add a bit of incidental detail. If you know what people ate, you know how they spent much of their time. The information can reveal a whole way of life. For instance, it has long been known that Iron Age communities on the west coast of Scotland kept cattle. But were they used for milk or beef? The answer makes a big difference to our understanding of that society.

"Compared with meat production," says Oliver Craig of the University of Newcastle, "dairy is a high-input, high-output, high-risk operation, indicating an intensive and sophisticated economy." In a paper published in Nature last November, Craig and others reported on a technique, known as digestion and capture immunoassay, for identifying proteins on ancient pottery. Used on shards from pots of the Scottish farmers, it revealed traces of milk.

While burning minute fragments of material from pots can tell you what was cooking, doing the same to hair can reveal remarkably precise details of diet. The considerable amount of hair that survives on mummified corpses and those lying in oxygen-free peat bogs is a mine of information that is only beginning to be exploited.

ÖйúAƬ

"Locked away in every strand of your hair is a record of your diet," says Steve Macko, an organic geochemist at the University of Virginia in Charlottesville. "Every meal lays down its chemical signature, and they survive for as long as the hair does." By analysing different isotopes of carbon nitrogen and sulphur, Macko can tell whether people were eating grains, meat, fish or vegetables.

Using this approach, he has discovered that the ancient Egyptian elite of the Middle Kingdom ate a restricted, largely meat diet, while Christian Copts from 8th-century Egypt had a diet with as much variety as you might find in a supermarket today. But he has walked into a storm of controversy over his findings on the diet of Otzi.

Macko concluded that the pattern of elements in the hair of this Neolithic man suggested he had been a vegetarian. It was a diagnosis that raised a few eyebrows because other studies had yielded the picture of a hunter. Not only was he carrying a powerful bow and long-distance arrows, but his tool-kit of knife, axe and scraper had clear signs of blood, hair and collagen.

Then, last year, a paper in the Philosophical Transactions of the Royal Society carried a report of an investigation of the iceman's stomach, which found signs of meat. The research by James Dickson of the University of Glasgow also analysed Otzi's hair and concluded that meat could have made up 30 per cent of his diet.

ÖйúAƬ

"The two findings are not mutually exclusive," Macko says. "For a portion of his life he was what we could call a vegan, but he was clearly opportunistic."

Literally putting flesh on long-dead bones and reconstructing the contents of dusty cooking pots promises not only to bring aspects of the past more vividly to life, but to change many long-held views.

ÖйúAƬ

McGovern's next project is to examine the residues from a pot found in a 9,000-year-old Chinese tomb. There has been some argument about whether rice wine came first in China or whether they were using grapes that long ago," he says. "I'd like to settle the debate."