What is the best of food?
Milk, for many a change comes out of it; butter and cheese are made of it, and it will feed a little child and an old man.
What is the worst of food?
Lean flesh. J.F. Campbell, West Highland Tales, III, 48.
The logic of this chapter is to some extent reversed. We know that dairying was of fundamental importance to the native population of Scotland: was it also a Mesolithic trait? The argument here is simplified; I have omitted the detailed objections to recent genetic proposals, which will be presented in a forthcoming volume. It is enough here to say that these genetic red herrings rely on the flawed Indo-European-Neolithic-Invasion thesis which is discussed in the first chapter. When a dramatic change of population is ruled out, it becomes clear that many developments now placed in the Neolithic are in fact much older. The use of milk is one of these developments.
The Scottish Highlanders, whose culture and language are a reliable touchstone for archaic survival, were devoted to milk. From the lexical evidence, summarised below, they were milkers before they came into Scotland in the Mesolithic. They traditionally divided their upland pastures into deer forest and summer pasture for cattle. Men hunted and women milked and the combination of summer milk, cheese and butter and winter venison was a fruitful one. Their tame cows bred with wild aurochs bulls, an animal known to Gaelic tradition as the uruisg or brownie and remembered as a rough but friendly creature, half-human, half-animal, who lurked around the winter homesteads and did good deeds, no doubt attracted by the tame house cows. Milkmaids lured him into the cow-pen with libations of milk.
The very early control of animals such as reindeer and aurochs in Western Europe has been suggested by a variety of writers, on the basis of evidence which is necessarily fragmentary, having survived against all the odds for twenty or thirty thousand years. One benefit of close control of wild or semi-domesticated animals is that the herders could milk the females. Campbell of Islay saw reindeer being milked by Lapps in a scene that could well be from the Palaeolithic.1
‘First, each of the girls took a coil of rope from about her neck and in a twinkling it was pitched over the horns of a hind. The noose was then slipped round the neck and a couple of turns of rope round the nose, and then the wild milkmaid set her foot on the halter and proceeded to milk the hind, into a round birch bowl with a handle. Sometimes she sat, at others she leant her head on the deer’s dark side, and knelt beside her. … The blear-eyed one, and the boy, and our party, went into the cota and dined on cold roast reiper and reindeer milk. … It was sweet and delicious, like thick cream.’
The milking routine at Gairloch towards the middle of the nineteenth century had an equally timeless quality.2 The estate then maintained a herd of sixty cows and their calves. The morning milking began at 6 am and took three hours to complete, and the whole performance was repeated in the evening. The milkmaids carried pails, three-legged stools and hobbles or buarach made of strong hair rope with a loop which fastened over a large wooden button made of rowan wood (to keep away the fairies). The calves had been penned separately overnight but were now released, one by one, as their mothers were being milked, so that the milkmaid on one side shared the milk with the calf on the other. The dairy equipment consisted of many flat, shallow, wooden dishes and wooden churns, casks and kegs, all of which had to be sterilised every day to keep the milk sweet. This was done by lighting a huge peat fire, heating hundreds of selected beach-stones or dornagan, the size of a fist, until they were red hot, and dropping them into the vessels. Three or four hot stones were enough to make the water boil immediately. Osgood does not say how many milkmaids were employed but it was clearly a full-time job for a large number of women, particularly if the preparation of butter and cheese are to be included.
If the Gaels are any guide, Adam was a hunter and Eve was a milk-maid. Place-names tell us that in settled areas the hill pastures were divided between deer and cattle, between deer forest and summer grazing, which ensured a supply of venison in the winter and dairy products in the summer (with cheese, when it was discovered, as a valuable supplement). In the Mesolithic both deer and the cattle were still wild animals, free to migrate as they had always done, while the humans who exploited them followed along. Deer were never domesticated - probably because there was never any reason to do so - but cattle were, several times in different places - probably because it is easier to milk a tame animal than a wild one. In recent times the migration to the hills took place on or shortly after 1 May (Beltane) while 1 November (Samhain or Hallowe'en) marked the start of the hunting season, when cattle had to return to the winter townships. In recent times these were fixed annual events marked by bonfires and other rituals, but transhumance may originally have been governed by the natural seasonal migration of wild cattle. Beltane in that case marks the start of the milking season, when the calves are born, while Samhain marks the end of the milking season, after which the cows were left to their own devices over the winter. The Irish milk festival, Imbolc, takes place on 2 February. This probably reflect the earlier births in a more temperate climate. Highland cattle living in England now give birth between 1 February and 30 April which fits the transhumance timetable very well. On 1 May the bulls are put in with the cows. Again, after Beltane the cows on the high grazing could mix freely with other wild cattle.
In the eighteenth century there are complaints about tenants’ cattle invading the deer-forests of Atholl: ‘the inference must be that in the eighteenth century shealings were considered more important than deer.’3 The 4th Duke of Atholl noted with disapproval that in 1774 ‘even what was denominated Forest was studded with such Sheals, and grasing of the deer and numbers quite reduced’.4 The atypical disturbance of the deer may tie in with the fact that after the Disarming Act of 1746 the tenants were no longer able to hunt deer and so had every reason to concentrate on cattle. The Duke appears to have followed the traditional arrangement when he restricted his tenants to defined areas which lay outwith the deer forests, but in less than a generation these high farms in Atholl had been abandoned.
The lexical evidence for milking
Prehistoric dairying can be deduced from both linguistic and genetic sources. The linguistic argument depends on the principle of convergence and divergence. This is explored in depth elsewhere but the basic principle can be deduced very easily by comparing American English and British English. As we all know, America was settled after c.1600 by English-speaking immigrants and remained a colony, in close contact with Britain, until 1775, when it declared its independence. From then until c.1950 America and Britain very largely they went their own way. Now again, since c.1950, there has been considerable contact but the two cultures, once identical, are now markedly different. So are their two languages. These differences are not random. A little research shows that the names for identical items in American English and British English are either convergent or divergent; they are either the same in both languages or they are different. A little more research confirms that items in use before c.1600 and up to 1775, such as door, horse, house, and window, are convergent; that these words were part of the English language used in both countries when America was still a colony. But almost all the names of items invented between 1775 and c.1950 (and since then) are divergent: flashlight for torch, trunk for boot of a car, sidewalk for pavement, railroad for railway, and so on. These items have two different names because when they came into use the two cultures were separated, by politics and by the Atlantic Ocean, and each made up its own new name.
The same argument can be carried back into prehistory. Convergent words identify elements of a common culture, as there was in the Palaeolithic, and divergent words show later evolution in separate communities, after permanent settlement in the Mesolithic. The only exception of any importance is when a valuable and unique item, such as copper, is traded from one place to the next with its name attached. This is a rare event. We know that trade in copper in Western Europe was closely controlled by Cypriots who sold it as 'the metal from Cyprus'. Slavs have their own word: we can deduce that they did not trade directly with Cyprus. But this way of naming, like the items so named, is rare and exceptional. There is nothing exceptional about the cow, which has been living in Europe for 250,000 years. When we look at names for the cow we find that most of the languages of Western Europe call a cow by the Q word such as E. cow or Du. kuh, or the alternative P word such as G. bo (the aspirates include Fr. vache) (for more details see below). From this we can deduce that the cow belongs to the earliest spread of European culture and language in the Palaeolithic. The only other possibility is that the cow arrived, like copper, as an entirely new animal in the Neolithic or Bronze Age, many thousands of years later. Some people believe this actually happened. Apart from the cave paintings, it is now confirmed that the aurochs was the genetic ancestor of the domesticated cow. The two are close enough to be regarded as the same species and have interbred in several places.5 In other words, a cow has always been a cow. Genetically and lexically there has always been a cow in Europe and it was always the same animal.
A second line of argument uses the same concept of convergence (common inheritance) and divergence (a later feature). Wild cattle, wild pigs and red deer were native to the whole of Europe (except perhaps Ireland) but sheep, goats, and probably horses came into Europe as domesticates in the Neolithic or Bronze Age.6 Since the cow, wild pig and deer were familiar to all the early occupants of Europe their European names ought to be convergent. The results of a little investigation are shown below. It also follows that European names for sheep, goat, tame pig and horse should vary locally, since they arrived after settlement and would be named locally as and when they arrived. This is again is what we find. This supports the proposal that a convergent name, such as 'cow', identifies an animal known all over Western Europe in the Palaeolithic, and that divergent names were invented locally in a piecemeal fashion for novel animals which arrived more recently. One might argue that the cow was merely an animal known to hunters, that milk could be human milk rather than cow or reindeer milk, but the table shows the convergent production of cheese and butter, which implies that cows were being milked in the Palaeolithic in Western Europe.
Table: Milking and other terms in Western Europe
C marks general convergence. D marks general divergence. Italics mark local divergence.
| English | Dutch | German | Danish | Gaelic | French | |
| ox | os | ochse | okse | (ur)uisg | boeuf | C |
| cow | koe | kuh | kue | bo | vache | C |
| calf | kalv | kalb | kalv | laogh | veau | C |
| milk | melk | milch | mælk | bliochd | lait | C |
| butter | boter | butter | smør | ìm | buerre | C |
| cheese | kaas | käse | ost | càis | fromage | C |
| cream | room | rahm | fløde | cé | crème | C |
| porker, boar | varke | ferkel | orne | uircean, orc | porc | C |
| pig, swine | big ‘piglet’ | schwein | gris, svin | muc | cochon | D |
| sheep | schaap | schaf | faar | aodh | mouton | D |
| horse | pfaard | pferd | hest | each | cheval | D |
This table shows that words for cattle and for dairy products are strongly convergent throughout most of Western Europe. The English terms were certainly in use before Britain became an island. The main divergence comes in words for butter and cheese in Danish and Gaelic. This is probably because the Danes and the Gaels live further north, were dependent for longer on reindeer, came late to dairying, and then evolved their own divergent terms.
The convergence of words for milk, butter and cheese over much of Western Europe allows us to place milking and the taming of cattle in the early Mesolithic or, more probably, in the late Palaeolithic. As far as I know archaeology cannot yet recognise milk, butter or cheese (or perhaps has never looked for it in such a context) but it can confirm that wild cattle lived over most of Europe in the company of the earliest settlers and that the aurochs cow is one of the most powerful of the beasts in the cave-paintings.
The logic of domestication
A useful principle is that man only domesticates species which are already useful to him. The question of function is central to the process of domestication. The main function of a herd of domestic cows is to produce milk. Meat and hides are useful by-products; the calves may be eaten; but in Scotland people hunted red deer to provide them with meat and hides. The life of the milkmaids on the summer pastures was purely pastoral. Women and children seldom if ever ate red meat.
Domestication never marks a new start but is always a means of securing a supply of some essential food resource. Milk was important as a food before any milk-animals were domesticated. The effort represented by the domestication of reindeer and, later, cattle, sheep, goats, camels, and horses, and the effort required to maintain breeding herds through dry summers and cold winters reflects the importance of securing a supply of fresh milk. Many nomadic tribes do very little else. But this great development could not have been made and would not even have been contemplated without considerable previous experience of handling milk-animals in the wild state. There is more to it but the reason which persuaded our ancestors to domesticate the aurochs was probably the convenience of having a tame milk-cow. Another pointer to a long period of prior exploitation is the fact that cattle were domesticated in several places independently, long before we detect any change in the species. Domestication in this case meant controlled breeding, excluding wild bulls from the cow pens.
The new improved Neolithic cow sooner or later replaced the wild cow all over Europe, either by direct improvement of the wild stock or by interbreeding with imported stock. The African zebu is thought to be evidence for long-distance trade in domesticated cattle between India and East Africa in prehistory. The purpose of the Neolithic upgrade was not to change the species but to maintain its survival. The function of the new cow was exactly the same as the function of the old cow: to produce milk. Our ancestors invested in new improved cows because they depended on a regular supply of fresh milk. To understand why we must look at lactase.
The lactase argument
Another argument for milking in Palaeolithic Europe rests on the distribution of the lactase enzyme.7 The lactase argument has recently been diverted by attempts to reconcile the facts with the Indo-European fallacy mentioned above, but the facts show Europeans as the heirs of a Palaeolithic milk-drinking population. They are the embodiment of the milkers implied by the convergent milking words. The argument is very simple.
Anyone who produces the lactase enzyme can digest the lactose sugar found in fresh milk. Those who do not produce lactase and who nevertheless drink fresh milk suffer persistent diarrhoea which upsets their electrolyte balance and which is eventually fatal. In most parts of the world, lactase is produced only by young children, but the majority of adult Europeans continue to produce lactase and so share the ability to digest fresh milk. As we would expect, the current distribution of the lactase enzyme matches very closely the spread of milk animals and a liking for fresh milk. Some neighbouring populations, such as the Turks, do not produce lactase but like the convenience of living off milk animals and have learned to ferment milk, a process which retains the food value and convenience of fresh milk but converts the lactase to lactic acid. Milk is not used in China, because very few Chinese produce the lactase enzyme after the age of three or four, which is the normal human pattern. After that age they find fresh milk, butter and cheese and those who eat them repellent. The pastoral animals which Europeans take for granted are found only in areas settled by people who are persistent lactase producers and who are therefore able to digest fresh milk as adults.
How did this genetic quirk spread through the European population? We are apparently presented with a chicken-and-egg logic, in which milk-animals would not have been domesticated without a general demand, and the general demand would not (could not) have existed without the presence of cooperative tame milk-animals. In fact two processes took place, more or less simultaneously.
The only way in which an advantageous gene can increase to 100 per cent in a population is for that population to be the result of positive selection for that gene in the form of a survival bottleneck. In the case of milk, the bottleneck excluded adults who could not tolerate fresh milk. This did not happen by accident or coincidence. Populations like the Turks use milk but do not produce lactase; conversely, the Chinese have a small proportion of persistent lactase producers but do not use milk. The Chinese represent the norm. But the ancestors of the bulk of the modern European population were the survivors of a bottleneck, or a series of bottlenecks, which favoured those who could digest fresh milk. This lactase event evidently took place very early in the settlement of Europe. That there were any survivors is due to the fact that every population has a small percentage of persistent lactase-producers - even in China. In a culture where fresh milk is never used (the normal situation) they are never exposed to milk as adults and their ability has no value and remains unrecognised. However this small number of adult lactase-producers - three or four out of every hundred - would be the only ones to survive a crisis where the only available food was milk. Only they would thrive on this abnormal diet. The majority of the population would die out. This is an all-or-none selection and it would take very little time to select for lactase. The death of most of the adult population over a few weeks or months would be followed (and evidently was followed) by the proliferation of a population selected for their ability to digest fresh milk as adults. There is no other plausible explanation for the distribution of the lactase gene in the European population.
This distribution tells us one more thing which also links with our milking words. Since the European population is still notably, if in places decreasingly, adapted to fresh milk, the survivors of the lactase bottleneck must have continued to use fresh milk as a staple food in every generation from perhaps 50,000 BP onwards. (Some day we may understand enough about genetics to give a reliable date for this event.) Since lactase production persisted, this dietary quirk also persisted. In Scotland most of the medieval population - women, children, old men and unemployed hunters - appear to have lived through the summer on milk. It is possible that only this sector of the population drank fresh milk but it is clear, from the persistence of LP levels approaching 100 per cent, that an ability to digest fresh milk after the age of three or four gave the individual a positive advantage. There are some signs of this. An over-reliance on dairy products can produce iron-deficiency anaemia, and this was diagnosed in three children aged from 5 to 10 who died c. 3300 BC and were buried in a single grave in the south of England, together with a woman aged 30 who was the mother of the youngest child. They had lived on a high-protein diet, probably based on milk as the adult had almost perfect teeth, and the evidence for iron deficiency suggests an absence of meat.8 Europeans do not realise that their milk-culture is a peculiar aberration but this is so. They claim that milk is the perfect food but fresh milk has no place in the diet of most or the population of Earth. Recent exposure to milk has not led to its universal adoption. The logic of this major division in diet places milking in the Palaeolithic, in Asia, among those who went on to settle Europe.
‘Since the right to keep or graze a cow applied to almost every class in the pre-enclosure farming communities, the milk products, fresh or preserved in some way, were the main defence against hunger. The cow was the form of social security of the period, as long as it was in milk.’ A. Fenton 1976, 156-7. Certain parts of Scotland produced so much cheese that it was used as a form of currency, paid in lieu of rent or instead of military service. ‘Both cheese and butter formed part of the teinds or tithes that had to be paid to the Church. The quantity of cheese paid as part of secular rents was such a regular feature, that the word kain, a payment in kind for rent, came to mean a certain quantity of cheese, about 60 cwt., in Argyll, Ayrshire, Dunbartonshire, and Galloway, and the dairyman who paid his rent in cheese was a kainer.’9 At the end of the day, of course, all this huge quantity of cheese was eaten, often spread with butter to make it more palatable.
The lactase story suggests that pastoral activities, as well as hunting, should show up in early European place-names and language. In practice pastoral place-names are difficult to recognise in Scotland, suggesting that rounding-up cattle to milk them used the same techniques and the same terminology as rounding up deer to kill them. We can recognise hunting terms with some confidence in the remoter parts of Highland Scotland, since large tracts of upland were never used for anything else, being too high and too wild for settlement. But township names, which were pastoral units, use a variety of topographical terms. A peculiarly Scottish word for a pastoral unit is dabhach, 'place of horned cattle'. Watson correctly noted that in Ireland a davach is 'a large tub with two handles' but in Scotland it means 'a large measure of land … reckoned to support so many head of stock’.10 These are in fact two distinct words. Kenneth Jackson's view (in his notes on the Book of Deer) of the davach as an arable unit is based on a misreading of Pennant. Pennant in his Tour of Scotland in 1772 said that there were things called davochs in Wester Ross but that it was impossible that they could be arable units since there was no arable land there to speak of, only tiny patches among the rocks. He also notes that a half-davach in Loch Broom supports sixty cows and their followers and is managed by eight men, eight women and an overseer, which seems conclusive. The best-preserved davachs are those which make up the parish of Strathavon in Banffshire, which has a good claim to be the wildest and wettest settlement land in Europe.
Another exception are the airidh-ruaidh names used for summer pasture in Perthshire11 They appear to be early as most shielings in this area are known by general topographical names. These peculiarly Scottish names may relate to a Bronze Age or Iron Age reorganisation, perhaps when modern cattle were introduced and carefully herded to keep them safe or prevent uncontrolled breeding. To the same date or to a similar reorganisation we can attribute the building of the crannogs (c.600 BC) and the so-far undated circular homesteads or cattle pens of northern Perthshire.
A further complication is that any search for early pastoral settlement must take account of the progressive loss, over the millenia, of coastal grazing in Scotland. The deer forests of the interior must have been disturbed more than once as displaced families and their herds retreated inland from drowned pastures on the coast. Inland areas would also become more attractive as the climate improved. The need to accommodate refugees may explain marginal inland settlements like Strathavon which consist of peripheral hunting forests and a string of low-lying townships or dabhachs whose summer grazing or shealings were within the deer forest at a short distance (five kilometres on average). Such a unit was largely defined by topography and was often identical to a fearann or parish.
Early Mesolithic Ireland was entirely different. It had wild boar but no native wild cattle or red deer, and this created a thinner pattern of settlement which relied much more on maritime resources. This is always a sign of marginal survival, in this case leading to an apparent extinction in the late Mesolithic, with repopulation from Britain (probably from south-west Scotland) shown by a radical change in the type of stone tools used. Did these new settlers bring red deer with them? Perhaps. But the population of Ireland became well-established only when domesticated cattle arrived, for Ireland, where the grass grows all year round, is the ideal country for cows. It is likely that pastoral farming spread very rapidly, creating a fairly uniform picture all over the country. In Ireland, unlike Scotland, there is a regular name for a township – baile. It appears to mean 'cow place'.
Appendix: Incidence of Lactase Persistence (LP)12
LOW: non-milkers, some with admixture of European genes.
China 1-7%; Japan 10%; Bantu Africans 11%; Inner Mongolians 12%; Australian Aborigines 15%; Lebanese 22%; Kazakhs (Xinjiang) 23%; Mexicans (rural population) 26%; Egypt 27%; Italy 28%
MEDIUM: admixtures of European genes, southern European populations adapted to cereals or an urban diet with little or no milk.
Southern France 35%;Italy (north) 48%; Italy (south) 59%; Saami (Russia and Finland) 40-75%; Balkan 45%; Greece 47%; France 50%; Spain 50%; American-Africans 55%; Central Asians 60%; Southeast Asians 62%; Poland 63%; Eskimos (Alaska) 64%; Portuguese 65%.
HIGH: milking populations who enjoy fresh milk though some are adapting to urban diets.
Bedouin 75%; Fulani 77%; India 80%; Tutsi 80%; American Asians 80% (?); Italy (centre) 81%; Finland 82%; Northern France 83%; Germany 85%; Eastern Slavs 85%; Austria 87%; Swiss 90%; England 95%; Australian Europeans 96%; Denmark 97%; Sweden 98%; Ireland 98%; Sweden 99%; Netherlands 99%.
There is currently interference (with high but undiagnosed morbidity among non-LP children) in marginal populations of American-Africans, American Asians, Alaskan Eskimos and African famine victims. The process of feeding a milk-based diet to non LP children will rapidly produce a high LP population, as predicated for prehistoric Europe.
