As calculus forms, it traps tiny particles of food within it. At the smallest level, these include starch granules – tight packages of stored sugar – and plant phytoliths – which are microscopic, silica-rich structures that help to provide support in living plants. In the lab, these particles can be analysed and identified. Calculus studies have revealed all manner of surprising details about ancient diets. Thanks to their dirty teeth, we now know that, forty-six thousand years ago, Neanderthals in what is now Iraq were eating cooked cereals, probably barley; that the Easter Islanders ate sweet potatoes; and that people in prehistoric Sudan ate a plant called purple nut sedge, regarded as a weed today.
This is all very well, but what about the presence of milk in human diets? There are no microfossils in milk – but there are highly characteristic molecules, and one of them proves to be an essential clue. It’s milk whey protein, or, more formally, β-Lactoglobulin – BLG. And importantly, for archaeologists, BLG is present in animal milk, but absent from human milk. It’s also relatively resistant to being destroyed by bacteria, so it tends to stick around for a long time. Another useful feature of this protein is that it varies between species – it’s possible to tell the difference between BLG from cattle, buffalo, sheep, goat and horse.
In 2014, an international team of researchers published their work looking for BLG in a range of archaeological samples. They found plenty of BLG in the calculus from Bronze Age teeth – going back to 3000 BCE, from both Europe and Russia – from cattle, sheep and goats – where there’s plenty of evidence for dairying, and none in Bronze Age teeth from West Africa, where there isn’t. So far, so good. This BLG study also cast some light on why Medieval Norse sites in Greenland were finally abandoned. Other studies – of nitrogen isotopes, no less – have suggested that, over 500 years – during a period of deteriorating climate – the Greenlandic Vikings were shifting to eat less food from domestic animals, and more from marine sources, including seals, before finally abandoning their settlements in the fifteenth century CE. Fish bones are often poorly preserved on archaeological sites, and it’s likely that the later Vikings were eating fish as well as seal. Rather than being pathologically inflexible about their diet, as the scientist and writer Jared Diamond suggested in his book Collapse, it seems the Greenland Vikings were trying to adapt. Whatever the reason for their abandonment of the Greenlandic colonies, it wasn’t an aversion to eating food from the sea.
The analysis of the calculus on the Viking teeth reveals another dietary change. At 1000 CE, the early Greenland Vikings were eating plenty of dairy products. But four centuries later, BLG had disappeared. So they weren’t eating domestic animals any longer, and they didn’t even have any access to dairy products. Perhaps the collapse of their dairy herds helped to speed the end for this Viking colony. But it’s also possible that the real reason for the abandonment of Greenland could have been more baldly economic. The Greenlandic Vikings traded walrus and narwhal ivory – but as supplies of African ivory started to enter the market, their goods were no longer so valuable. Time to leave this place, then, as the bottom fell out of the ivory market and where you could no longer even get a nice bit of cheese.
All of this is fascinating, and the newly unlocked potential for using β-Lactoglobulin to reconstruct ancient diets is exciting – but this latest study looked no further back in time than the Bronze Age. Very soon, I imagine someone will go looking for milk whey protein in more ancient teeth, and I’d like to think that very faint traces might pop up even before the advent of domestication and dairying in the Neolithic, amongst the unbrushed teeth of our hunter-gatherer forebears.
Potsherds and cowherds
Our ancestors didn’t pay much attention to brushing their teeth, and it seems they weren’t that keen on washing up, either. To date, the earliest definite evidence we have of humans drinking milk comes from the fatty residues left on the insides of ancient sherds of pots from the Near East, dating to the sixth and seventh millennia BCE. A team headed up by Richard Evershed at the University of Bristol looked at 2,225 potsherds from south-eastern Europe, Anatolia and the Levant. They found a hotspot of early milk use close to the Sea of Marmara. This study of milk and pottery drags us away from the Fertile Crescent, to the greener and lusher north-western corner of Anatolia. And it made a lot of sense: Neolithic sites in this area contain high proportions of domesticated cattle bones, and this is an area with high rainfall and lush pastures compared with most of the Middle East. The bones tell a story of their own – with plenty of young animals in the archaeological assemblages, the early farmers seem to have been rearing cattle for both meat and milk.
The results of this study of ancient potsherds seem obvious in a way, but until Evershed and his colleagues found those traces of milk fats, it was thought that dairying was a relatively late addition to the Neolithic mode of life, coming along several millennia after the original domestication of animals, and perhaps two millennia after pottery was invented. The new evidence pushes dairying right back – to the same time as the appearance of the very earliest ceramic vessels in western Asia, in the seventh millennium BCE. Is this more than a coincidence? Perhaps needing something in which to store and process milk could even have prompted the invention of pottery.
Nonetheless the earliest evidence for both milk – and pottery – still comes along some two millennia later than the earliest appearance of domestic animals, including cattle, sheep and goats, in the ninth millennium BCE. And with these techniques, clever as they are, it’s impossible to know if milk was used any earlier – because you’re then in a world before pottery and there are simply no potsherds for milk lipids to stick to.
Another frustration with the evidence from milk lipids on pottery is that, unlike milk whey protein in calculus, this time we don’t know which of the possible animals the milk came from – it could have been from sheep, goats or cattle. It may be possible, however, to get to the bottom of that by carefully studying animal bones from Neolithic sites, and an investigation carried out across eleven archaeological sites in the Central Balkans did just that. Analysis of cattle bones from those sites revealed an increasing proportion of adult animals over time. On average, adult animals made up only about 25 per cent of cattle bones in Neolithic assemblages. When numbers of young cattle are high, this suggests a focus on rearing animals for meat. At later, Bronze Age sites, from 2500 BCE onwards, 50 per cent of cattle bones come from adults. The shift towards older animals suggests that ‘secondary products’, such as milk (and perhaps traction as well), are becoming more important. The pattern seen in sheep bones is similar. If this pattern is reflected elsewhere, this would suggest that the first cattle and sheep were domesticated for their meat, and milking of these species came along later. But the goat bones in this Balkan study revealed something different. A higher proportion of adult animals was seen right from the beginning of the Neolithic – which starts around 6000 BCE in the Balkans – suggesting that herders in this region had always exploited domesticated goats for their milk as well as their meat. As soon as they had goats, they had goats’ milk.
And yet some other recently published research prompts us to be cautious about generalising from that Balkan study. There’s good evidence from other sites that cows’ milk was being used right back in the early Neolithic. Once again the clues come from potsherds. This time, it’s all about cheese. The first step towards making cheese involves getting particles of a specific milk protein, casein, to start sticking to each other, creating a protein net which traps fat globules inside it. This mess of coagulated protein and fat is the curds. What’s left behind is a thin fluid containing some soluble proteins – the whey. There are two main ways in which you can transform milk into curds and whey: you could acidify the milk, or you could add an enzyme to it – usually rennet. Heating the milk can also speed up the process.
All of this must have been discovered by accident, by Neolithic farmers trying out new recipes perhaps, or even new storage solutions. Just
imagine that you’re a Neolithic farmer, off herding your animals for the day, and you want to take some milk with you. Pottery is great – but a bit heavy to carry around. Instead, you decide to use a bag made from a goat’s stomach. It’s not such an odd idea – bags like this are often used for water. Anyway, you fill it up with milk and off you go. Later in the day, you go to take a sip of milk and something strange has happened – it’s become watery, with lumps. The rennet – the enzymes sticking around inside the goat’s stomach – has transformed the milk. Rather than throw it away, you take it home and show your family. They’re all quite impressed by this brand-new dairy product. But it gets even better. If you can separate the curds from the whey, you’ve got the beginnings of cheese. You could use a cheesecloth, or a metal sieve. Neolithic people may well have used cheesecloths, or even wicker sieves; though, unsurprisingly, neither have been discovered at any archaeological site. Cloth is not generally the sort of stuff that stands the test of time. And, in the Neolithic, metal sieves were still a long way off. But there are plenty of examples of perforated pots, which have been widely interpreted as cheese strainers. Some people have suggested other uses for these pots – ranging from lamps, to honey-straining, to beer-making. Richard Evershed’s team turned their attention to fifty fragments of perforated pots from Neolithic sites in Poland, dating back to as early as 5200 BCE.
They detected lipid residues on 40 per cent of these pottery sieve sherds. And in all but one, the lipids were identifiable as milk fat. It was proof of the cheese-strainer theory – and the first definite evidence of prehistoric cheese. By processing milk, these ancient people had also done the lab scientists a favour: fresh milk residues don’t last long on pottery – but the fats change when milk is processed, and persist much longer. And at these archaeological sites in Poland, 80 per cent of the animal bones are from cattle. While the milk lipids could have come from cows, goats or sheep, it seems most likely that the Neolithic farmers of Poland were indeed milking cows, and making cheese from their milk. The domesticated aurochs was here to stay.
Bones and genes
The earliest archaeological evidence of domesticated cattle themselves comes in the form of bones from a pre-pottery Neolithic site called Dja’de-el-Mughara, right on the banks of the Euphrates River. It’s an extraordinary site – an ancient farming village which later became used as a graveyard in the Bronze Age. Deep down in the Neolithic layers, there are a few human burials, but also carved bone ornaments, a large circular building with wall paintings – and the butchered bones of the animals that these early farmers were keeping. There, around the Euphrates, the rolling grassy plains would have provided perfect pasture for early domestic herds during spring and winter. During the parched summer months, the villagers could have driven their animals down to the river’s edge, or even out on to islands, just as they still do today. From the tricky task of managing wild herds – just think of those horns – to capturing a few aurochsen and breeding from them, the farmers had started the process of domestication. Compared with aurochs, the bones of domesticated cattle are smaller, and there’s less difference between males and females. There’s also a difference in horn shape, which is reflected in the bony horn core that projects from the skull. This early, skeletal evidence of cattle dates back to between 10,800 and 10,300 years ago, around the same time that the first firm evidence of cereal domestication appears in the Levant. Sheep and goats, though, are thought to have been domesticated a little earlier – perhaps just a few centuries before. It seems to make sense that domestication of these animals began to happen before the domestication of crops really took off. Pastoralism – tending herds of animals – is almost a halfway house between a nomadic, hunter-gatherer lifestyle, and a settled, agricultural way of life. But the transition from hunting and gathering to pastoralism could be very swift. One site in Turkey, A¸sikli Höyük, shows a change from people subsisting on a diet including a wide range of wild animals to one where sheep made up 90 per cent of the animals eaten, over just a few centuries. Whatever prompted the pre-pottery Neolithic people at A¸sikli Höyük to manage those flocks of sheep, they effectively ended up with a way of storing meat – creating a walking larder – that made their food sources more reliable.
Early genetic studies suggested that sheep and goats were domesticated many times, in separate places, but all broadly within south-west Asia. In fact, it’s more likely that there was a single centre of domestication for each species – but then plenty of interbreeding with wild cousins. Domestic goats come from the wild goat, Capra aegagrus, while sheep are the domesticated descendants of the wild sheep or Asiatic mouflon, Ovis orientalis. The European mouflon, on the other hand, appears to be a domestic breed turned feral, rather than an ancestor.
It looks like a similar story for cattle. For a long time, it was believed that the two main subspecies of domestic cattle – taurine and indicine – came from separate origins. Darwin certainly thought that may have been the case, writing in the Origin, ‘I should think … that [the humped Indian cattle] … had descended from a different aboriginal stock from our European cattle.’ And to be fair, Bos taurus indicis, also known as zebu cattle, do look quite distinct from Bos taurus taurus, or taurine cattle. Zebu cattle have a large hump above their shoulders, and a long dewlap hanging down between the front legs. They’re also much better suited to hot, dry conditions than taurine cattle. Studies of mitochondrial DNA and Y chromosomes supported this idea of a separate origin for each subspecies. But a single origin makes much more sense: it seems most likely that domesticated cattle arose in the Near East, between 10,000 and 11,000 years ago, and then spread, meeting wild relatives along the way. Reaching South Asia some nine millennia ago, a significant level of interbreeding with local aurochsen could have introduced zebu genes and characteristics to domestic cattle.
The diaspora of cattle got under way very quickly. Farmers and their cattle were travelling west too; by 10,000 years ago, someone had been brave enough to put them in a boat and take them to Cyprus. By 8,500 years ago, domestic cattle had reached Italy, and by 7,000 years ago, they had spread, with the early farmers, into western, central and northern Europe, as well as Africa. Cattle had reached north-east Asia by 5,000 years ago. As sheep and goats spread out from the Middle East, they were moving into uncharted territory for caprines – there were simply no wild relatives to interbreed with. But it was different for the domesticated bovines: wild oxen ranged right across Europe and Asia, and cattle seem to have interbred with them everywhere. The first clue came from mitochondrial DNA, where unusual variants in Neolithic cattle from Slovakia, Bronze Age cattle from Spain, and in a few modern cattle as well, were all traced back to European aurochsen. More recent genome-wide analyses have revealed widespread interbreeding between domestic cattle and local, wild oxen, right across Europe. British and Irish cattle breeds, in particular, have a lot of aurochs DNA in their genomes. But we can surely only speculate on how deliberate – from a human perspective – any interbreeding may have been.
I’ve spent some time living with indigenous reindeer herders in Siberia, where the domestic reindeer herds are large and impossible to guard or corral. The wild herds are even larger, and – like the domestic herds – often on the move. The reindeer herders I spoke to worried less about wild animals joining their herds, than about losing their reindeer to the untamed hordes. They were always nervous when they knew there was a wild herd nearby. Their experience has made me think differently about those early farmers and their herds.
Just how carefully did Neolithic farmers tend their cattle? Did they fence them in or let them roam more freely? Did they catch and add carefully selected wild aurochs to their herds, or does the genetic introgression simply record unavoidable contact between domestic and wild animals? If this is the case – and I have no idea if it is – then this simply means that aurochs cows were more likely to join up with domestic herds than wild bulls were.
From a biological perspective, it’s not su
rprising that domestic cattle continued to interbreed with wild populations. The two modern subspecies of cattle have often interbred to produce hybrids. In Africa, cattle DNA reveals a history of male zebu cattle being bred into herds of taurine cattle, to produce Sanga cattle. In China, taurine cattle spread into the north and indicine cattle into the south. This north–south divide is still evident in Chinese cattle today, with taurine–indicine hybrids in the middle. Cattle can also produce hybrids with other species. One Chinese cattle breed has been found to contain yak DNA – and, conversely, domestic yaks contain DNA from cattle. In Indonesia, zebu cattle often interbreed with the local wild cattle species, known as banteng, or Bos javanicus.
The riddle of the shrinking cow
When they entered into an alliance with humans, cattle, sheep, goats and pigs changed. In contrast to grains of wheat, which grew larger under domestication, cattle and other animals got smaller. But curiously, cattle – unlike sheep, goats and pigs – then continued shrinking, through the Neolithic, Bronze Age and into the Iron Age. And it was a significant reduction. Archaeologists have been able to quantify the shrinking that took place during just the Neolithic by scrutinising ancient bones from European cattle, where farming got started around 7,500 years ago (5,500 BCE). By the end of the Neolithic, 3,000 years later, cattle were, on average, a third smaller than they’d been at the beginning of farming.
Tamed Page 12