How to Track a Mammoth

Liam Shaw

From a guide to the fossil mammals and birds in the department of geology and palaeontology in the British Museum (1896)

A male mammoth died 17,100 years ago, during the last ice age, alone above the Arctic Circle. It was 28 years old. Thousands of years later, one of its tusks was found near the headwaters of the Qikitaġruraq, a gravel stream in Alaska. In Science last month, researchers described how they had used measurements taken from the tusk to plot the mammoth’s wanderings over the course of its life.

We call them mammoths because of their tusks: the name is probably derived from the Mansi languages of Western Siberia, meaning ‘earth-horn’. For centuries these horns were occasionally dug up in Russia, but were of unclear provenance. The earliest quotation in the OED entry for mammoth is from Richard James’s 1618 Dictionariolum Russico-Anglicum: ‘Maimanto, as they say a sea elephant, which is never seene, but accordinge to the Samγites he workes himself under grownde and so they finde his teeth or hornes or bones in Pechore and Nova Zemla’. (James travelled to Russia as the English ambassador’s chaplain.)

These gigantic underground remnants of ‘never seen’ animals led the French naturalist Georges Cuvier to the theory of extinction. In 1796 he argued that mammoths were different from modern-day elephants – a species which had been irrevocably lost.

Until twelve thousand years ago, mammoths roamed the northern hemisphere on the ‘mammoth steppe’, a huge expanse of tundra which stretched across Eurasia from Britain to Alaska. Deep excavations in London often uncover mammoth remains: in 2016, a jawbone fragment was found under Canary Wharf during work on Crossrail. What little we know about individual mammoths is often constrained to their last moments, based on where their skeletons were found – a struggle in the sticky tar – or the record of violent trauma inflicted by human weapons. But recent developments in isotope dating allow for longer narratives.

Like trees, tusks have growth rings (though unlike trees, they grow from the base). Each layer in the Qikitaġruraq mammoth’s tusk was made using atoms it had recently ingested from the plants it consumed: tough grass, sedges, mosses. One of the trace elements is strontium, which has four naturally occurring isotopes. The ratio of strontium-87 to strontium-86 in the soil varies slightly, but measurably, from place to place. This ratio is passed on to the plants that grow in the soil, and therefore in the teeth and bones of the animals that eat the plants. The different strontium ratios in different parts of the mammoth’s tusk offer a forensic clue to where it was at different times in its life.

The tusk investigated by Matthew Wooller and his colleagues was almost two metres long and more than twenty centimetres in diameter at its thickest. It had to be split open lengthwise. First, the researchers traced a thin line with a diamond-coated circular saw, just a centimetre deep. They then used this as a guide for a band saw. A small wheel in the base of the saw ran along the incision, keeping the tusk aligned as the blade cleaved through the 17,000-year-old ivory. It took six people to guide the tusk, one of them inserting spacers to stop it closing up and snapping the blade.

With the tusk’s interior exposed, the team took more than 340,000 strontium ratio measurements, around thirty for each day of the mammoth’s life. Averaging over hundreds of measurements – a timescale of around a week – produced a smooth function, a record of the mammoth’s motion in strontium-space.

To map this onto the earth’s surface, the researchers divided the Alaskan region into a grid of square kilometres. The strontium ratio in each square, known from geological surveys, could be matched to the data from the tusk. The starting point was the only square definitely known to have been occupied by the mammoth: the spot where it died. Every other location had to be inferred, one place at a time, in reverse order.

Each new position was determined according to the strontium isotope ratio, distance from the previous position and glaciation level (mammoths couldn’t travel across large areas of ice). The closer the geological strontium ratio was to the corresponding tusk measurement, the more likely it was the mammoth would have been there.

Stringing the selected points together produced a reverse-walk across the Arctic landscape, as the virtual mammoth trudged back in time towards its birth. Thousands of these 28-year itineraries were computed, each a possible journey the real mammoth might have taken. Individually, every trajectory was a meticulous fiction – the exact true path remains unknown – but overlaying them revealed the most likely route taken by the mammoth. (The researchers also used related measurements from an oxygen isotope to give greater confidence.)

The animal spent most of its early life in the lower Yukon River basin. In its juvenile years it made regular north-south movements, probably as part of a migrating herd. Then, at the age of 16, it set out on its own, perhaps in much the same way that male elephants today tend to leave female-led herds at reproductive maturity. In total it walked nearly fifty thousand miles. For its last two winters, its range dwindled to a small region north of the Brooks Mountains. The researchers hypothesise that it died from starvation, in late winter or spring.

‘The attention of Cuvier has been much fixed on fossil bones,’ the Edinburgh Review observed in 1811; ‘and he has extracted from thence, by his profound skill in comparative anatomy, much curious and precise information concerning the ancient inhabitants of the globe.’ Using methods Cuvier could not have dreamed of, that work continues.


  • 17 September 2021 at 1:18pm
    John Morgan retired says:
    Very fascinating, thanks.

  • 21 September 2021 at 7:17pm
    Paladin says:
    Here's a link to the original article (warning - NOT FREE!):