In Defence of Dogs

by John Bradshaw

  In short, when dogs first pick up a trace of an odour they are trying to locate, they can only guess roughly where it is coming from. This is where their wet noses are helpful. Technically known as the rhinarium, the ‘pebbled’ area of specialized skin around the nostrils is crammed with pressure and temperature sensors. (The rest of the skin on their heads is generally rather insensitive.) Those wet noses are not only directly sensitive to the wind going by, but are also simultaneously cooled on the upwind side, giving them an instant readout of which direction to head in.

  Yet because of the wind’s intrinsic vagaries, this tactic will almost certainly immediately take the dog out of the scent pocket. Inexperienced dogs will immediately switch to quartering, running mainly across the wind in an attempt to relocate the odour. As soon as they pick up the smell again, they will check their pace and revert to running upwind, until the scent is almost inevitably lost again, whereupon the whole process starts again. More experienced dogs, for instance those that have regularly competed in gundog trials, will tend to go on running upwind for a few seconds after they have lost the odour, confidently expecting that they will soon run into another pocket of odour. If they do not, they will then switch to the quartering tactic. A prolonged period of running without encountering any odour will usually cause most dogs either to give up, or, if their motivation remains strong, to start a purposeful loop back downwind to the general area where they first detected the odour.

  Once dogs get close to the source of the odour they change their tactics. The average strength of the odour will be increasing at this point, but the dogs will probably be alerted more reliably by the sudden disappearance of the gaps of clean air between the pockets of odour. Now they will slow down abruptly, tails wagging furiously, and start to use their eyes to locate the target, because the olfactory information is no longer sufficiently detailed to be of much help. Only if they accidentally blunder on past the source and hence lose the scent will the dogs loop back around into the odour corridor and go back to using their noses.

  Whether tracking or following an air trail, dogs can ramp up the sensitivity of their noses. Rapid sniffing makes the air entering the nose more turbulent, so that more of it comes into contact with the olfactory membranes. They can also change the airflow in their nostrils: by widening the nasal valve, dogs can send more information into the olfactory area. Dogs that are tracking along the ground need to travel quite slowly in order to maintain contact with the trail, and so can sniff all the time, at about six sniffs each second. They can also temporarily increase their rate of sniffing if they need to, up to twenty sniffs for each inhalation. Indeed, they may even be able to perform the saxophonist’s trick of breathing in through their noses, sniffing continuously, while simultaneously breathing out through their mouths. But dogs cannot be sniffing all the time, particularly while running – they would simply run out of air. When attempting to follow an air trail, as opposed to a track on the ground, dogs often move at a canter. At such a pace, they of course use up a great deal of oxygen, which requires them to breathe more heavily. Because of that, they have to make a trade-off between maximizing the amount of odour they can analyse and locating their target quickly. When gundogs follow a track laid on the ground, they sniff five or six times a second. However, they sniff only about twice per second when following a scent upwind, and even less frequently – once per second or so – when running crosswind trying to locate the scent plume.

  Once inhaled, the contents of each sniff are analysed by the dog’s olfactory system. The air in each sniff passes up the nasal passages and swirls around the turbinates, scroll-shaped bones that carry the olfactory receptors. The receptors encode the nature and strength of the odour, and then pass on this information to the olfactory nerves and thence to the brain, where the sensation is generated and comparisons can be made with odours that have been sampled in the past.

  Until recently, not much was known about how mammals detect and analyse odour, but now that the canine and other genomes have been sequenced, an increasingly detailed picture is emerging. The molecules that make up the odour – and there will be many different types in any natural odour – first have to be extracted from the air and passed on to the receptors. Because odour molecules move very slowly, the receptors need to be very close to the air; otherwise, it would take so long for the odour molecules to reach them that they would not be able to give the instantaneous response a dog needs in order to keep up with its six sniffs a second. Thus in the dog’s nose, the receptors are just a few thousandths of a millimetre from the open air. Exposed like this, the receptors are very susceptible to damage. The olfactory equipment protects itself in two ways. First, it cleans, warms and humidifies the incoming air by passing it over mucus-covered membranes before it can reach the olfactory epithelium itself. Second, the olfactory neurons, on which the receptors sit, are constantly renewed, undergoing replacement roughly every month.

  Once absorbed by the olfactory epithelium, a specific odour triggers the corresponding receptors. The odour molecules diffuse through the mucus that covers the receptors’ exposed ends. In the dog, there are hundreds of types of olfactory receptors. All those of a particular type are connected to a single ball of nervous tissue about a tenth of a millimetre across, which in turn relays its information to the brain through a small number of nerves, the mitral cells. This combining of information maximizes the capacity of the nose to detect tiny amounts of odour. In contrast to the eye, what matters is not where in the nose each molecule is being picked up, but how many receptors are being activated at any one time, so it makes sense for all the signals to be gathered together before relaying the information to the brain.

  The olfactory bulb in the brain then compares all the signals produced by specific receptors to generate much more nuanced information – much as our brains enable us to ‘see’ millions of colours even though our eyes can detect only three. Clearly, since dogs can distinguish tens of thousands of different odour molecules, even though they have only 800 or so types of receptor, there cannot be one receptor for each odour. Rather, a particular odour molecule binds to a greater or lesser extent to several different kinds of receptor and the ‘odour brain’, the olfactory bulb, combines the information to decode what has been detected. Finally, once the odour molecules have interacted with a receptor, special enzymes degrade them quickly, otherwise the sensation of the odour would persist for too long. The receptor is then clean and ready to receive the next incoming molecule.

  Dogs have yet another way of perceiving odours, and it is one that we humans do not share at all. Running between their nostrils and the roof of their mouths, just behind the front teeth, are a pair of fluid-filled tubes, the incisive ducts. From each of these runs a cigar-shaped, blind-ended tube called the vomeronasal or Jacobson’s organ. If you look behind your own incisors in the mirror, you will see nothing – we do not have incisive ducts or functioning vomeronasal organs, although most other mammals (and reptiles) do. Like much of our olfactory sense in general, these organs disappeared in our distant ancestors, way back during the evolution of the higher primates.7

  The purpose of the vomeronasal organ (VNO) is not easy to determine. Both the VNO and its ducts are fluid-filled, so at first glance they would seem to be rather awkwardly placed to perceive odours. However, there is a muscular pump that can move the fluid in and out of the nose and down into the VNO, providing a possible mechanism for odour molecules to get from the outside world into the VNO. Potentially, scent molecules can first be absorbed into the saliva, or into fluid in the nostrils, and then be pumped to the VNO. The resultant delay means that the VNO cannot be much use for detecting information that changes by the second, such as an airborne scent-trail, but should be adequate for an animal to analyse the scent of another member of its own species, which should remain constant. Thus the VNO is thought to be the sense organ that specializes in social odours, although not exclusively so: the roles of the VNO and the nose overlap considerabl
y in this respect.

  It is still not clear precisely what dogs use their VNO for. Part of the problem is that there is no obvious external sign that indicates when a dog’s VNO is being brought into play. Cats and some other mammals – including the dog’s close relative the coyote – make a characteristic facial expression when employing the VNO: the mouth is held slightly open and the upper lip is curled back, and the cat/coyote seems momentarily lost in thought. This expression is exhibited not when the animal is sniffing food – food odours are analysed by the nose in cats and coyotes – but when it is sniffing a scent-mark left by another member of the same species. Although dogs do not display the same expression, some dogs do chatter their teeth when they are sniffing scent-marks, and others make a kind of chortling noise. These sounds may be indicators that the pump that transfers odours into the VNO is in use.

  The canine VNO thus certainly does something. Most likely it is used to analyse social odours, which the dog either licks up or inhales, simultaneously activating the pump that carries them to the VNO itself. Odours can thus be analysed twice – first, more or less instantaneously, by the nose and then for a second time, in a more leisurely fashion, by the VNO. The highly detailed information that emerges can then be stored away in the brain for use in future social encounters. This is yet another example of the dog’s superiority over humankind in its ability to decode smelly information – but one that has thus far proved difficult for scientists to decipher, probably because we have difficulty appreciating a sense that we do not possess ourselves.

  Whether they are perceived through the nose or the VNO, smells are very important to dogs, much more so than they are to us. Dogs use odour not just to decide what to eat and what not to: odour is their primary means of identifying people, places and other dogs. Smell is their dominant sense, the one they use in preference to all their other senses, whenever they can.

  Because odours are so complex, and because they differ depending upon the environment in which the individual dog lives, it would be impossible for dogs to be born with the ability to recognize more than a handful of odours. Dogs therefore have to learn what each odour means. They start learning how to use their sense of smell even before birth, in much the same way that human babies learn the sound of their mother’s voice while still in the womb.

  Using ultrasound, scientists have observed puppies exercising their breathing muscles in the womb during the two weeks before they are born. This ‘breathing’ almost certainly allows the puppy to learn something about its mother’s characteristic odour, including the type of food that she is eating most. In one experiment, pregnant bitches were given food flavoured with aniseed throughout the last three weeks before they whelped.8 As early as a quarter of an hour after they were born, and before they had even begun to suckle, their puppies moved towards the smell of aniseed. The smell of vanilla, an odour that the puppies had never been exposed to, did not have the same effect, so clearly they were not simply investigating the aniseed because it was an unfamiliar smell. What had caused this preference for aniseed? Presumably it had flavoured the mother’s amniotic fluid, where the pups had ‘smelled’ it before birth.

  It is not clear precisely why puppies need to learn their mother’s smell before they are born; it would probably be equally helpful if they learned it immediately after birth, since at that stage they are in any case too helpless to move far from their mother. (Prenatal learning is especially useful in animals such as sheep, where the young are much more mobile at birth, and thus run the risk of getting separated from their mothers.) Perhaps dogs have simply retained this ability from their mammalian ancestors, even though it is not of much use to them now.

  From the moment they are born, puppies use odours to help them make sense of the world around them. Initially, this attention is focused primarily on the mother. Three days after she gives birth, the bitch produces a substance around her mammary glands that, in turn, is modified by bacteria on her skin to create an odour which helps her puppies locate her. It also seems to have a calming effect on them. Although the mechanism is somewhat obscure, the same substance appears to have a calming effect on adult dogs. Scientists are not agreed as to whether this change in behaviour is also caused by an instinctive, ‘pheromonal’ effect, or is due to a memory of being protected by the mother, but the extracted odour has proved useful in the treatment of acute fears in dogs.

  As soon as they are able to move around, dogs start sniffing anything and everything they come across, a behaviour that continues for the rest of their lives. Many owners are embarrassed by their dogs’ proclivity for sniffing the crotch of every person they meet; many more train their dogs not to do this. Yet this is the dog’s first-choice method for identifying other animals, both human and canine. When two dogs meet in the park, their first and often only goal is to sniff each other. Sometimes they circle one another before diving in for a sniff. Sometimes one dog is so intent on sniffing the other that a chase ensues. However, in eight times out of ten the objective of the encounter is to get olfactory information about the other dog.9

  Because both dogs attempt to be first to sniff the other, they often end up circling one another

  The odours in question evidently originate at both ends of the body. Sniffing is often focused around the ears, indicating that these may be the source of an individual-specific odour. But although the ears definitely contain odour-producing glands, little is known about what kind of information is contained in the odours they produce. Sniffing under the tail must pick up odours from the preputial (male) and vaginal (female) glands that add their contributions to the dog’s ubiquitous urine-marks.

  However, the main target of canine sniffing seems to be the anal sacs. Located on both sides of the anus, as their name implies, these contain a pungent mixture of odours (mainly produced by micro-organisms) that varies considerably from one individual dog to another. Perhaps because the anal sacs are usually closed, their odours do not vary much from one week to the next, though they do change gradually over timescales of a few months or so.10 They are therefore good candidates for a ‘signature’ odour, albeit one that will, like all chemical signals produced by mammals, require relearning by recipients as it changes gradually over time. This is presumably part of the explanation for dogs’ insistent attempts to sniff the back ends of every other dog they meet; if the odour stayed consistent from one month to the next, they would need to do it only very occasionally. Such sniffing, like urine-marking, has origins that go back to well before domestication. Young male wolves have a similar fascination with sniffing this part of the body, and adult wolves sometimes actively invite other members of the pack to sniff them in this area, by standing stock-still with their tails held upright.

  Scent inspections tend to follow predictable patterns. Some dogs, mostly males, go straight for the area under the tail, which produces an information-rich odour. Most females, and a few males, prefer to sniff the head of the other dog first, and then to move back, provided the other dog will let them. This difference between the sexes is so far unexplained, but it is also true of wolves, so it may simply be an inherited tendency with little functional significance as far as modern dogs are concerned.

  Interestingly, while dogs love to sniff other dogs, it seems that most dogs do not much like being sniffed themselves. It is almost always the dog being sniffed that attempts to break off the interaction. While dogs want to find out as much as they can about other dogs’ ‘odour signatures’, they seem reluctant to give away their own. (Young wolves share this reluctance, so this may be where this behaviour originated.) It is as if they see information about the dogs around them as the key to – well, something – since once the sniffing is finished the interaction usually ends. What that ‘something’ is, has yet to be resolved. If they are successful in their encounters, dogs go home from walks with lots of information in their heads about what the other dogs in their neighbourhood smell like. What they do with this information is currently unclear.<
br />
  In fact, there is a great deal we do not know about the kinds of information dogs can get by sniffing each other. The anal sac may signify more than individual identity; for example, for a wolf it might also indicate which pack it belongs to, if members of a pack share an odour. Some components of the wolf’s anal sacs may also vary according to gender and reproductive state. The same might be true for dogs, but at present we do not know. It is remarkable that we know so little about the one activity, sniffing, that dogs like to do the most. Nothing better exemplifies how human-centred we can be when thinking about our domestic animals. Somehow we fail to grasp the ‘otherness’ of much of what they experience. Of course, to dogs what something smells like is not ‘other’; it is, if anything, more important than what it looks like.

  The dog’s fascination with odour must have originated way back in its evolutionary past. In particular, scent is a major mode of communication for a wide variety of animals (it is humans who are the exception, not dogs). Scent is a good way of transmitting information between animals that live far apart from one another. The early carnivores, the dog’s remote ancestors, are very unlikely to have lived in groups. They were almost certainly solitary, defending territories against other members of their own kind. The only groups would have been mothers and their dependent young, who would have stayed together for a few months at most before the young were old enough to disperse. Communication between adults would therefore have revolved around establishing and maintaining territorial boundaries. Apart from courtship and mating, face-to-face meetings would have been a rarity. Not only that, they would also have been risky: being well-armed with teeth and claws, carnivores try to avoid disputes that damage both parties, not just the loser. Finally, these animals were probably nocturnal, inhibiting visual communication. In the natural world, all these issues can be circumvented by using scent-marking as the primary mode of long-distance communication. A scent-mark designed for the purpose can last for days. Messages can be left for recipients to pick up at some undetermined moment in the future, obviating any necessity for actual meetings to take place. Contemporary dogs, which have evolved from sociable animals and have become yet more sociable with domestication, may no longer need to scent-mark as frequently as they obviously think they do; but their wild ancestors must have found it very advantageous, and their legacy remains in our dogs’ everyday behaviour.