In Defence of Dogs

by John Bradshaw

  While research into the origins of modern dogs is far from over, the available evidence makes it clear that comparisons between domesticated dogs and American timber wolves should be treated with caution, to say the least. With one proviso: the research has so far been focused almost entirely on maternal DNA; therefore, the descendants of a mating between a male American wolf and a female domestic dog would not be picked up. Some American dogs may therefore carry American wolf genes from one or more long-distant male ancestors; research should be able to resolve this soon. What is clear, however, is that no female American timber wolves were successfully domesticated. It is impossible to determine, thousands of years after the event, whether the reason for this was that American wolves were intrinsically difficult to domesticate, or that the first human colonizers of the Americas, having brought their own dogs with them from Asia, saw no need for further domestication and rather, being hunters, would have perceived the local wolves as competitors. Whatever the explanation, the fact remains that the vast majority of domestic dogs are only very distantly related to the American timber wolf, separated as they are by 100,000 years of evolution. This is the first reason why comparisons between wolf behaviour and dog behaviour need to be treated carefully: most studies have been done on a type of wolf that, if it was any more distantly related to domestic dogs, would probably be considered a different species.

  Scepticism about comparisons between wolves and dogs is further warranted by the fact that, although the DNA analysis indicates that dogs descended from Eurasian grey wolves, none of the wolves that have been studied over the past seventy years or so, American or European, can possibly be considered the ancestors of the domestic dog: the two certainly had a common ancestor many thousands of years ago, but there is no evidence to suggest that modern wolves closely resemble these common ancestors. Indeed, logic dictates precisely the opposite.

  A simplified representation of the evolution of the wolf and the domestic dog. Genetic ‘bottlenecks’ occurred: at each of the domestications (one failed domestication is shown); when dogs were divided into breeds; and when many local wolf populations went extinct, leaving the survivors isolated from one another in different parts of the world.

  Wild wolves, as they exist today, are almost certainly quite different in behaviour from their – and dogs’ – ancestors. As soon as agriculture began in earnest, all wolves that had not been domesticated would have inevitably become a threat to the newly formed herds of livestock, and so they would have been persecuted by humans. Until firearms became widely available in the eighteenth century, human efforts to eradicate wolves were fairly ineffective, but thereafter it became possible to exterminate wolves from whole areas. For example, the wolf population in Norway and Sweden declined dramatically from the 1840s onwards; the DNA of today’s Scandinavian wolves shows that they are descended from immigrant Russian animals, which had hung on in isolated areas throughout the twentieth century, until the modern conservation movement gave them the space to move into areas recently vacated by their cousins. This pattern of local extermination, or at least severe population reduction, was repeated throughout Europe; the few wolves that survived must inevitably have been those that were the most wary of people. Today’s wolves are therefore the descendants of the wildest of the wild, whereas today’s dogs must be derived from a much more tameable sort of wolf, one that is no longer found in the wild and about which we know almost nothing.

  So, although the DNA of dogs tells us that they are, indisputably, wolves, much of the scientific study of wolf behaviour conducted in the twentieth century must now be regarded as of dubious significance to our conception of dog behaviour. Most of the wolves that were studied were kept under conditions that we now know were not only highly artificial, but also highly stressful for many of the individuals concerned, thereby inducing highly abnormal behaviour. Furthermore, we have no reason to suppose that any of these wolves precisely resemble the type of wolf that was originally domesticated. Such wolves no longer exist; they are far from extinct, since they have millions of living descendants – our dogs – but they no longer exist in the wild. Their nearest wild equivalents are probably feral dogs, domestic dogs that have reverted to a wild or semi-wild existence; these, too, are at the same time the direct descendants of the wild ancestor of the dog, and they live a similar, independent, lifestyle.

  Since comparisons with the wolf are no longer as valid as they seemed as recently as a decade ago, my approach is to widen the search for the biological characteristics that make up the dog’s true nature. Some of the characteristics that facilitated domestication may indeed be much more ancient even than the wolf, going back many millions of years to long-extinct species that were themselves the ancestors of wolves, jackals and wild dogs. All of these current species have features in common, features they most likely share with their common ancestors: they live in family groups, where the young adults often help their parents to raise the next round of offspring, have excellent noses, are highly intelligent and adaptable, and are either hunters, or scavengers, or both.

  Ultimately, there may be nothing special about the wolf that singled it out for domestication; perhaps it just happened to be the social canid that was in the right place at the right time.8 Unfortunately both that place and that time are currently shrouded in mystery, but what is certain is that the aberrant and atypical behaviour of modern, captive wolves is highly unlikely to be of any value in understanding either the behaviour of these ancestral wolves or that of domestic dogs. Rather than focusing exclusively on the grey wolf, we should regard the dog as a canid whose closest living relative happens to be a wolf. It is the possession of the canid toolkit that was vital to the successful domestication of the dog – a story whose roots are intimately twined with those of our own.

  2

  How Wolves Became Dogs

  The story of the dog’s domestication – its evolution from wolf to its own unique sub-species of canid – parallels that of our own emergence into civilization, from the hunter-gatherers of the Mesolithic through to the modern age. There were domestic dogs well before any other animal was domesticated, so arguably the dog is likely to be more altered relative to its ancestors than any other species of animal on earth. The process of domestication has stripped away much of the detail of the ancestral species, but dogs nonetheless retain some of the characteristics of the more ancient lineage that gave rise to jackal, coyote and wolf alike. Dogs are somewhat like each one of these, but they are also unique, the only fully domesticated canid, and much of what makes them unique was introduced by that very process of domestication. The story of that domestication therefore makes an essential contribution towards understanding what our dogs are, and what they are not.

  Over the last decade, we have learned a great deal about the domestication of the dog. The sequencing of the DNA of hundreds of individual dogs has forced a reappraisal of all the previous evidence regarding the process of domestication. While there are undoubtedly more surprises to come, the broad scope of how it happened, as well as much of the detail, is now fairly well established.

  In addition, we have new perspectives on when and where the dog may have been domesticated. We can be reasonably sure that there were several, possibly many, attempts at domestication of the grey wolf in various parts of the world, but that ultimately the products of some of those domestication events – in places other than North America – ultimately endured, whereas others did not. The process of discovery is still ongoing; ancient bones and fossils that were formerly identified as unequivocally belonging to wolves are being re-examined, in case they might actually have come from early wolf-like dogs. The evidence has been clear enough, however, to place the separation between wolf and dog further back in their evolution, by thousands of years, giving more time for wolf and dog to diverge, an analysis that further undermines the idea that the behaviour of the dog is simply a subset of that of the wolf.

  While we are gaining a better understanding of t
he ways in which dogs are different from wolves, we are also learning more about how we, as humans, have helped to make the dog different. The domestication of dogs has been revealed as a complex process, more convoluted than that of any other animal, leading not only to radical changes in the shape and size of their physical bodies, but also to an almost complete reorganization of their behaviour. Furthermore, while mankind has guided that process, it is only in the past century and a half, and only in the West, that we have taken control of it completely. For 10,000 years or more, as the purposes for which dogs were valued changed and proliferated, dogs have coexisted and co-evolved with us. Essentially, they domesticated themselves as much as we domesticated them.

  When was the dog domesticated? Until fifteen years ago, the answer was thought to be simple. The oldest remains of dogs found by archaeologists were carbon-dated to be no more than 12,000 years old, 14,000 at most. This timeline placed the first dogs before the beginnings of agriculture about 10,000 years ago, and well before the domestication of any other animal. So the dog was, for this reason alone, considered a special case: the pioneer for all subsequent domestications, such as goats, sheep, cattle and pigs. Because the dog was domesticated so early in the history of humans, there is little detailed evidence as to how wolves became dogs, a paucity of information that has left a great deal of room for speculation as to why and where this first occurred. Until fifteen years ago, however, at least the ‘when’ seemed well established; no bones had been found that had both unequivocally belonged to a dog and were more than 14,000 years old, so domestication must have occurred no earlier than about 15,000 years ago.

  Then, in 1997, a team of scientists from the United States and Sweden made an astonishing claim: they had sequenced DNA from living dogs and wolves, and the findings indicated that domestication could have taken place more than 100,000 years earlier.1 If this was true, it would mean that dogs were man’s companions not just at the birth of agriculture, but right at the dawn of our own species – as soon as modern humans had emerged from Africa, where they had evolved, and encountered grey wolves (which do not occur in Africa) for the first time. This announcement triggered a minor epidemic of speculation about the possible co-evolution of man and dog. Most archaeologists rejected the idea, pointing to the complete lack of dog remains that could be dated any earlier than 14,000 years ago. But there was nothing intrinsically wrong with the DNA data, even though its interpretation was still open to debate. Dogs, it seemed, joined us during our pre-agricultural origins.

  Since 1997, there has been a steady flow of more detailed studies of dog and wolf DNA, and, as a result of these, our conclusions about the exact moment of the dog’s domestication have changed, and are still changing today. DNA technology is relatively new, and while it may give unequivocal answers when used for ‘fingerprinting’ (for example when confirming the parentage of a particular puppy in a dispute over pedigrees), its use in reconstructing events long past is much more open to interpretation. Different types of DNA can give different answers; for example, that contained in the nucleus of most mammals’ cells (the subcellular organelle where paternal and maternal DNA mix) often tells a different story to DNA from other parts of the cell, such as the mitochondria (which contains only maternal DNA). As new analyses have appeared and been integrated into the picture, the estimate that dogs might have been domesticated more than 100,000 years ago has since been revised down considerably – now to between 15,000 and 25,000 years ago.

  One reason for this drastic downward revision is that problems have been found in the method used to calculate how much time has elapsed since two animals had a common ancestor. The DNA most commonly used for this purpose comes not from the nucleus but from the mitochondria (whose genetic content is abbreviated as mtDNA). Very occasionally – only once every few thousand years – mitochondrial DNA mutates, so that mother and daughter, who would otherwise have identical mtDNA, exhibit sequences that differ at just one location (this applies only to mothers; fathers do not pass on any mtDNA to their offspring, male or female). Unlike other kinds of mutation, these changes have no effect on the health or fecundity of the animal, and so are passed on ‘silently’ down the generations, spreading throughout all the daughter’s descendants. By counting the number of differences between two individual animals’ mtDNA, scientists can estimate the amount of time the two individuals’ lineages have been diverging – and can thus form an idea of how long ago their most recent shared female ancestor lived. The bigger the number of distinct mutations, the older the two animals’ joint lineage must be.

  Errors slip into this sort of mtDNA dating when scientists, having determined how many unshared genetic mutations exist between two individuals, attempt to work out just how often these mutations may have occurred in both animals. The regularity of these mutations varies from one kind of animal to another. However, scientists know from the fossil record and from carbon dating that the dog’s ancestor, the wolf, diverged from the coyote about a million years ago. A simple comparison between the number of differences between dog and wolf, and the number between wolf and coyote, suggests that the dog and the wolf have been separated for about one-tenth of that time – in other words for about 100,000 years. This calculation, however, relies on the mutations in mtDNA occurring at the same rate in domestic and wild animals. Since the 1997 study, it has become apparent that mtDNA mutations occur more frequently in domesticated animals than wild ones. The same comparative method used in the 1997 study has consistently overestimated the time since domestication for virtually every animal to which it has been applied: for example, the DNA of the pig, probably first domesticated 9,000 years ago, suggests a domestication of between 500,000 and 60,000 years ago; that of the horse, more than 300,000 years ago instead of about 6,000. The mutation rate must therefore be faster during domestication than in the wild, speeding up the rate at which mtDNA changes from once every few thousand years to once every few hundred. Studies of other species suggest that this accelerated rate may be a side effect of chronically high levels of stress hormones, caused by living in crowded conditions and in close proximity with man. Thus the estimate of 100,000+ years is highly likely to be an overestimate, perhaps by a factor of five or more, bringing the interval since the dog’s domestication down to a much more realistic 20,000 years or so.

  In addition to comparing the dog’s DNA with that of the wolf, scientists can also examine how much variation there is between different types of dog, as a way of determining how long they have been around. However, this procedure, too, superficially seems to suggest that dogs were domesticated much earlier than 20,000 years ago. A recent analysis of the DNA that codes for the dog’s immune system has produced an estimate of several hundred thousand years since domestication, a figure even more unlikely than the 100,000 years indicated by the mtDNA, since it predates the evolution of our own species. On the other hand, such an estimate assumes that mutation is the only source of variation, and that all dogs are descended from a single pair of wolves. A similar degree of diversity could occur if, say, several wolves had been domesticated, each of which had distinctive DNA. But this is likely to be the case only if each of those wolves had lived in a different part of the world, a supposition which in turn implies several domestication events.

  The apparent contradictions between the archaeological evidence and DNA evidence can be reconciled if we posit not just one domestication event, but several, in different parts of the world. It is now becoming possible to examine the DNA of fossilized dog teeth taken from Neolithic burial sites. While only a few dozen individuals have been sequenced so far, the results tend to confirm that wolves were indeed domesticated at several, possibly many, locations.

  Scientists have also begun to find proof for multiple domestications by looking at a different type of DNA, extracted from living dogs. The DNA that codes for the immune system is inherited from both parents, not just the mother, as mtDNA is. The much greater diversity in the DNA for the immune system su
ggests that dogs have far more forefathers than foremothers; in other words, dogs overall seem to have many male wolf ancestors between them, but only a few female wolf ancestors. Thus the genetic material from the ‘extra’ males must have been introduced after domestication had started. The early domestic dog bitches would presumably have been attractive to, and so occasionally mated by, wild male wolves. Moreover, their puppies would have been born in close proximity to humans, and, provided the genetic contribution of their wolf father did not make them too intractable, they could have survived to contribute to the dog genome. There is no reason why a mating between a male dog and a female wolf should not also produce puppies, but they would be born in the wild, and hence would be more likely to contribute to the wolf’s genome than to the dog’s.

  Thanks to recent scientific developments, we now know that the diversity of the modern dog’s genome is not hopelessly incompatible with the archaeological evidence surrounding the dog’s domestication. Nevertheless, there is still a discrepancy – possibly as large as 5,000 to 10,000 years – between the most likely date suggested by the DNA (20,000 or more years ago) and the oldest date that most archaeologists will agree to (14,000 years). The reason for this discrepancy probably lies in the type of evidence that archaeologists will accept as evidence for domestication. Human remains and the bones of wolves have been found together at sites going back half a million years, long before modern humans evolved, but archaeologists do not accept these joint burials alone as signs of domestication. Rather, they look for evidence of domestication either in the remains of animals that are clearly distinguishable from wolves (for example, those having a wider skull, a shorter muzzle or smaller teeth), or in signs that the animals, even if otherwise indistinguishable from wolves, had a special place in human society – and preferably both.