In Defence of Dogs

by John Bradshaw

  The key difference between a dog and a wolf is not what it looks like but how it behaves, and especially how it behaves towards people. DNA and bones cannot tell us how these early dogs behaved, or what their everyday interactions with people were actually like. Domestication affects outward appearance, certainly, but at the very earliest stages this is incidental. What defines an animal such as a dog is what goes on under the skin – specifically, how the behaviour of its ancestors has been altered to enable it to live comfortably in man-made environments.

  Although we know a great deal about the behaviour of today’s American timber wolves, and an increasing amount about the relict wolf populations of Europe, this information gives us little insight into the behaviour of the first domestic dogs. Modern wolves are only very distantly related to the domestic dog, and they have been under intensive selection pressure, especially over the past few hundred years, from those who wished to exterminate them. It is therefore unsurprising that today’s wolves are very difficult to socialize, and that tame wolves tend to remain unpredictable and potentially aggressive towards people throughout their lives. By persecuting wolves, we have selected those individuals that are naturally wary of us. It is therefore very difficult to derive any knowledge about early dogs from what we know about contemporary wolves. Moreover, this means that we cannot even replicate domestication, by taking wolves out of the wild and selectively breeding them to become more like dogs. Since the wolves that were the direct ancestor of domestic dogs are, in their original form, extinct, that would be impossible.

  One recent modification of a canid is widely held to provide pointers as to how wolves might have changed into dogs. This is the silver fox, a colour variety of the wild red fox, that is bred in fur-farms. Silver foxes are usually kept in cages and are scarcely tame, let alone domesticated, but in the 1950s a group of Russian scientists began to breed them selectively, using only the tamest individuals in each generation.8 At first, few of the foxes could be handled, even by a person offering a tasty food treat. After a few generations of breeding only from individuals that would tolerate handling, however, some individuals emerged that would actively seek contact from people. Indeed, after thirty-five generations, most of the foxes were behaving in a remarkably dog-like way – wagging their tails, whimpering to attract attention, sniffing and licking their handlers’ hands and faces. Some were even taken home as pets by the staff, who reported that these animals could be as obedient and loyal as domestic dogs. The geneticists’ objective of producing a fox that was easier to handle seemed to have improved its welfare too. Freed from the relentless fear and anxiety of having to encounter an alien species (us!) every day of their lives, the new ‘tame’ farm-foxes exhibit levels of stress hormones four times lower than those in the original ‘wild’ version. A similar reduction in reactivity, and susceptibility to stress, is evident when dogs are compared with wolves – a reduction traceable to changes in the hypothalamus, a part of the brain that is, among several functions, concerned with emotional reactivity. Such changes are probably a direct consequence of selection for tameness, so in this respect the tame farm-foxes may well be similar to the wolves that adapted to living near, and scavenging from, human settlements.

  The most interesting finding of the Siberian fox experiment was that the farm-foxes became easier to tame because the period before they became frightened of new experiences was effectively lengthened. Most young mammals go through a period of their lives in which they are naturally inquisitive and trusting. And it is usually during this stage that they are still being looked after by their parents, who are on hand to make sure that this characteristic does not get them into trouble. As they get older and more independent, the offspring become much more suspicious of anything unusual, and much more likely to run away after an initial inspection. In the farm-foxes, selection for tameness corresponded with an extension of this ‘trusting’ phase, which ends when wild foxes are about six weeks old, but lasts for about nine weeks in the ‘tame’ variety. That extra three weeks is enough to allow regular handling to take effect, producing a fox that trusts, rather than fears, the people who look after it.

  Another finding of the Siberian experiment has been used to posit the effect of domestication upon canids’ appearance, though this is largely unsubstantiated. The appearance of some of the tame foxes produced in the experiment is different from the wild variety; a few, though by no means all, of the tame foxes have unusual dog-like features, such as curly tails, floppy ears and white patches on their coats. Some authorities have claimed that such features are part and parcel of domestication, that selection for tameness inevitably brings with it all these changes in appearance. Unfortunately the data do not really support this idea. True, more ‘tame’ foxes have floppy ears than do the ‘wild’ ones, but they are still in a tiny minority – fewer than a quarter of 1 per cent. Fewer than one in ten of the tame foxes has a curly tail. Fewer than 15 per cent have a white ‘star’ on their forehead. Exactly how these changes became slightly more common in the ‘tame’ foxes is still something of a mystery, but they are still rare, and probably tell us little or nothing about domestication.

  While the Siberian experiment produced tame foxes, there is a significant difference between these foxes and domestic dogs in terms of the extent to which they are – or, it seems, can be – ‘domesticated’. In dogs, the process of acclimatizing to humans does not disrupt normal social relations with other dogs. By contrast, when the foxes develop a relationship with humans they seem to lose interest in socializing with other foxes. Red foxes – the same species as the farm-fox – are rather sociable animals, often living in groups of four to six animals. Yet the tame farm-foxes are solitary animals; as devoted as dogs, but as independent as cats. This contrasts with the domestic dog (and the domestic cat), whose social relationships can and indeed normally do develop simultaneously both with humans and with members of their own species (and potentially other species as well).

  If the tame foxes can tell us anything useful about the dog, it is that tameness, while a useful first step, is not the same thing as full domestication. Tameness permits the replacement of one set of social responses – directed at members of the same species – with another – directed at humans. Dogs, by contrast, need to retain both, in order to continue functioning as members of their own species while simultaneously establishing and maintaining relationships with their human owners. Nothing in the farm-fox experiment sheds any light on how this capacity might have come about during the domestication of the dog.

  The farm-fox experiment does show that selection for tameness can be extremely rapid – indeed, it seems to be fast enough to suggest a plausible first stage in the domestication of the wolf. The key difference between the two animals is, of course, that the foxes were a captive, isolated population that was deliberately selected for tameness. The wolves that were sufficiently tolerant of humans, on the other hand, selected themselves to be the ancestors of domestic dogs: those that were easily tamed could start breeding in the proximity of humans; those that could not, rejoined the wild population. The appearance of dog-like behaviour in the tame foxes, such as licking humans’ faces and hands, and whimpering, also supports the idea that the dog’s social repertoire is drawn not from that of the wolf exclusively, but rather from an ancestral palette of possibilities inherited from the canids as a whole.

  The farm-foxes tell us that natural variation in tameness within a species can be sufficient, in at least one of the canids, to produce individuals that could be the ancestors of a domestic animal. This experiment thus provides us with a model for the initial separation between wild wolves and those that were naturally tame enough to live alongside people. The resources that the naturally tame wolves were able to obtain from humans must have been sufficient to allow them to adopt a new way of breeding. Instead of hiding them away in a den, the intrinsically ‘tame’ mother wolves must somehow have allowed their cubs access to humans, so that taming, and selection fo
r tameness in subsequent generations, could proceed further. Underlying the onset of tameness are changes in the production of and reactivity to stress hormones, alterations that are evident from tame farm-fox and dog alike. However, the farm-foxes tell us nothing about the way that dogs gained the capacity to sustain social relationships simultaneously with their own species and with ourselves. Nor do they tell us anything about how dogs achieved their remarkable diversity of shapes and sizes – and yet this very diversity permits another, very different, approach to understanding the subsequent phases in the domestication of the dog, once tameable wolves had begun their association with mankind.

  Instead of comparing dogs with wolves, or trying to reconstruct the domestication process, we can find important information about how dogs came to be by examining the differences between breeds and types of modern dogs. The ways in which they differ from one another can provide clues as to how those changes in appearance might have come about. Since different-sized dogs appeared very early in the history of domestication, at least 10,000 years ago, it is possible that the processes that led to the diversification in body shape are the very same as those that permitted domestication to proceed beyond tameness. And since many of the differences between breeds and types of dogs are known to arise through alterations in the rates at which the body and behaviour develop in early life (alterations that are reflected both in the outward appearance of the dog, and in the way its behaviour is organized), the emergence of these superficial differences is arguably the most important underlying process that has produced today’s dogs.

  Dogs come in so many shapes and sizes that they have long been a puzzle to zoologists, but in fact many of the changes can be accounted for simply by a common biological mechanism, the technical term for which is neotenization. Roughly speaking, this refers to the phenomenon whereby growth in some parts of the body stops, while other parts continue to grow at the normal rate. If the whole skeleton stops growing earlier than usual, but the internal organs continue to mature, then the result is a smaller-than-usual animal that is still capable of reproducing. For example, the skeleton of an adult Lhasa apso is similar to that of a Great Dane puppy, but the Great Dane will continue to grow for many more months before it becomes sexually mature. If the growth of the skeleton is altered selectively, the end result is a change in shape as well as a reduction in size. The skull of an adult Pekinese has essentially the same proportions as that of a wolf foetus, but its body is more dog-like. In ‘toy’ dogs, the growth of the whole skeleton stops at what would be, for a wolf, a very early stage. In flat-faced dogs, the growth of parts of the skull is slowed to maintain the proportions of that of a foetal wolf.

  We are now coming to understand even more about the physiology that underlies these differences in canid appearance. It turns out that the skull and skeleton of the wolf change shape dramatically between their genesis in the foetus and their final form in the adult, under the control of various hormones. Much of the size variation in today’s dogs probably comes about through changes in the growth stages during which these hormones are produced, how much is produced, and how effective they are at doing their job. Thanks to all the work that is going on to unravel the canine genome, it should soon be possible to identify how these changes work.

  The very same principle of selective arrested development that governs dogs’ growth can also be used to explain how domestication moulded the dog’s behaviour. For example, dogs continue to play even when they are adult, unlike most animals. Because the behaviour of juvenile wolves is more flexible than that of the adults, the dog has been likened to a wolf that has never grown up, except in the important sense that it becomes sexually mature and so can reproduce. Its behavioural development has, in a sense, been arrested. The farm-fox story sheds important light on this process by telling us that tameable wolves probably differed from untameable wolves in having a delayed period of social learning at the beginning of their lives, such that tolerance of human contact had time to develop. Dogs, for their part, are like tameable wolves in which development of behaviour has been slowed down further still, to the point that it becomes arrested at the (wolf’s) juvenile stage, where behaviour is more flexible and can therefore be adapted much more easily to the requirements of humans. Some simple resetting of the dials that control the development of brains and behaviour can, in theory, account for much of the transition from wild wolf to tame, from tame to domestic, and then to the diversification of dogs into types of different sizes and shapes.

  One further difference between dogs and wolves can also be accounted for by a selective change in the development in the two animals: dogs become sexually mature somewhat earlier than wolves. Dogs are also fertile throughout the year, unlike wolves, which are sexually active mainly in the winter, leading up to the birth of the cubs in the spring. Both of these differences are likely to be consequences of the transition from the wild, with its seasonal but predictable food supply, to early human societies where food was more plentiful on average but also more unpredictable; proto-dogs that could breed any time after their first birthdays would have out-competed those that waited, as wolves do, until their second winter.

  For the same reason that they need to be much more opportunistic in grasping opportunities for breeding, dogs are also much less choosy than wolves in their choice of sexual partners. This is evident from the Y-chromosome (paternal) DNA of today’s dogs, which is much less diverse than their mitochondrial (maternal) DNA. Because wolves pair-bond, males and females are about equally likely to contribute to the DNA of the next generation. Given the promiscuous tendencies of male dogs, some males can potentially sire more than a hundred litters in their lifetime, while many others leave no offspring at all. Bitches are constrained by the fact that they can produce only one litter per year. Moreover, the variability in male reproductive success appears to have been set up well before the creation of the modern breeds in the nineteenth century, suggesting that male promiscuity is an ancient, not a recent, trait of dogs.

  The promiscuity of the male dog must have been one of the factors that helped man, first accidentally but then increasingly deliberately, to impose his own selection pressures on the species. Some of these choices might have been simply fanciful, such as a preference for a particular coat colour, or an especially ‘cute’ face – qualities of no particular consequence for the process of domestication. Other aspects of human behaviour – such as taking special care of the offspring of a bitch prized for her trainability and loyalty – might have pushed the process of domestication along.

  At the early stages of domestication, certainly up to the point that dogs became physically distinct from wolves, human intervention in breeding is unlikely to have been a conscious process, and indeed may have been haphazard, as it remains to this day in village dogs. The archaeological record indicates that dogs may have disappeared entirely from some societies, only to be replaced hundreds of years later by immigrants from elsewhere. Other societies may have rejected dogs even when they were available: although Japan was first colonized by mankind about 18,000 years ago, there is no record of dogs in that country until about 10,000 years ago – presumably Japan’s new inhabitants considered the dogs available in ancient China unsuitable, for some now impenetrable reason.

  Despite the almost certain lack of early deliberately applied selective pressures from humans, over the course of several thousand years wolves must have made some kind of faltering progress towards becoming an animal that had many of the behavioural characteristics of today’s dogs, even if it still looked much like a wolf. Certain physical changes, however, would almost certainly have begun to occur during this time. Dependence on man for what was probably a rather erratic supply of food would have favoured a reduction in body size. As dogs were transported into warmer climates, those with shorter, paler coats would have out-competed those with the wolf’s long, darker fur, producing a conformation that survives in village dogs to this day.

  Many of the oth
er conformations that we see in today’s dogs are also ancient. By 10,000 years ago, dog-keeping and therefore dogs themselves had spread throughout much of Europe, Asia, Africa and the Americas; soon after this, and in many parts of the world, recognizably distinct types of dog appear. Over the next couple of thousand years, dogs diversified rapidly, so that by the time representational art became commonplace some 5,000 years ago, there were already dogs for many purposes. Long-limbed, long-nosed sighthounds, superficially similar to the modern saluki or greyhound, were used for hunting.9 Heavy, large-headed mastiff types were used for guarding and general intimidation. Hounds were developed that hunted mainly by scent, suited to finding and following large game in thick cover. Subsequently, larger dogs were found to be useful as pack animals, either carrying loads on their backs or – as widely practised by some Native Americans – pulling a travois. Small terrier-like dogs were used for keeping rats and mice at bay, and for hunting animals that go to ground, such as rabbits and badgers. Lapdogs, similar to today’s Maltese dog, are first recorded from Rome more than 2,000 years ago, but there were probably already lapdogs in China by this time, possibly the direct ancestors of today’s Pekinese and pug. The arrival of lapdogs completed the process of generating the dog’s remarkable variation in size; any that became smaller, or larger, would probably not have been biologically viable in the days before veterinary care. Lapdogs were also the first dogs bred solely for companionship, though for many centuries these through-and-through pets would have been rare compared to dogs kept for more utilitarian purposes.