For dogs, like most mammals, it is more important to see all the time rather than to see particularly well, in order to remain vigilant to danger. We humans have sacrificed some of our ability to see in the dark in order to be able to see in colour and in great detail during the day; presumably our tree-dwelling ancestors had few nocturnal predators. For this reason, dogs’ eyes are adapted to see much better in the half-dark than we do, but less well (though perfectly adequately) in bright light.
In order to be more efficient at night, dogs’ eyes contain a structure that ours do not. Most dog owners who walk their dogs at night will have noticed that their eyes shine when a light is pointed at them. This is due to a reflective layer of cells behind the retina called the tapetum, which almost doubles the sensitivity of the eyes in low light. Furthermore, dogs’ eyes are connected to their brains differently from ours. We have a staggering 1.2 million nerve fibres in our optic nerves, which allows us to resolve a lot of detail, provided there is enough light. Dogs have a mere 160,000 connections, and, unlike our optic nerves, theirs are connected to multiple rods or cones, and can be triggered if any one of these receives a scrap of light. This enables dogs to see at lower light levels than us, but their perception of detail is inevitably reduced, by a factor of about four, since their brains have no way of knowing which particular one of a bundle of light-sensitive cells has been triggered. Put another way, perfect vision in humans is described as 20:20 whereas dogs can manage only 20:80 at best. Some may have even worse eyesight than this – it has proved difficult to design an eye test for dogs that is as sensitive as those opticians use on us. (Wolves, incidentally, have rather clearer vision than dogs; it is possible that dogs are descended from a more nocturnal wolf than those that survive today.)
Dogs’ eyes also produce a wider picture than ours, a picture that is less centred on straight-ahead vision. They can see more of their surroundings without moving their heads. The average dog’s field of view is about 240 degrees, compared to our 180 degrees – so that, unlike humans, they can see a limited amount of what is going on behind them. Each eye points about ten degrees off the centre-line of the muzzle, so there is a considerable area of overlap to the front, and dogs do use this to produce true binocular vision. Some breeds probably have better binocular vision than others; indeed, the degree of overlap is less in breeds with long muzzles.
But while dogs’ field of vision is larger than ours, their close vision is not nearly as good. Most dogs cannot focus much closer than 12–20 inches from their noses, even when they are young (though, as with humans, that minimum tends to increase as they get older). Once they get their snouts within a foot or so of anything they are interested in, other senses take over, especially their keen sense of smell.
The most notable difference between canine vision and our own, however, lies in the limited range of colours that dogs can see. Like most mammals, they have only two types of colour-sensitive cells (cones) and so can see only two primary colours – blue-violet and yellowish-green.1 Of course, like all mammals they can distinguish many different colours based on the relative strength of the signals coming from these two, but the absence of the yellow cone (which humans have) means that they cannot distinguish red from orange, or orange from yellow. There is also a gap between the colours that their two types of cones are sensitive to, such that dogs see turquoise as grey. Nevertheless, scientists who have measured dogs’ colour vision found that they were more attentive to colour than most other mammalian species (apart from ourselves and our trichromatic relations, of course), so it is likely that dogs do sometimes use colour in their everyday lives.
Thus, for instance, dogs running around in a park in the daylight will see much of what we see, with some small differences. They see more, because they can see to the sides of their heads as well as straight ahead; but also less, in the sense that the leaves on the trees and the grass will be rather similar muted shades of greyish-green, and red and yellow flowers will be similar to each other in colour. As the dog’s carnivorous ancestors did not need to pick out the ripest fruits or the most tender leaves, the latter deficiencies were most probably of no particular consequence to them, and probably matter little to dogs today. As darkness falls, however, the dog’s superior night vision comes into play, enabling dogs to continue running about happily in the undergrowth long after their owners need a torch to find their path.
Despite these minor differences, the degree of overlap between the dog’s visual world and ours means that misunderstandings of what they can and cannot see rarely cause problems. Anyone trying to train dogs to respond to visual cues based on their colour would do well to avoid using both red and orange, but this suggestion is unlikely to apply to more than a handful of specialist trainers; most everyday training uses sound and movement cues. Their general lack of interest in colour probably explains why few dogs display much interest in watching television – though the poor quality of television sound (as far as they are concerned) may also play a part.
Our restricted hearing, compared to that of dogs, can lead to situations where they are inconvenienced or, at worst, even suffer. Dogs’ hearing is significantly more sensitive, and more versatile, than ours. Their low-frequency hearing has a range similar to ours, but they can hear high-pitched sounds that we cannot hear at all. We refer to such frequencies as ‘ultrasound’ – although, if they could, dogs would describe us as having high-frequency deafness. Cats, which can hear even higher-pitched sounds than dogs can, would presumably describe dogs as having high-frequency deafness.2
It is easy to forget that dogs can hear sounds that we cannot. Some researchers into canine responses to sounds have used ordinary audio playback equipment, apparently unaware that this is designed to mimic just what we humans hear and does not reproduce the high frequencies that are presumably an important part of all sounds as far as dogs are concerned. Unsurprisingly, therefore, dogs sometimes react to ‘live’ sounds but do not necessarily react to recordings or broadcasts (television, for example) that are, to our ears, virtually identical. The equivalent experience for us would presumably be something like the difference between FM radio and long-wave AM (which does not reproduce high frequencies).
It is not clear why dogs (or wolves) would ever have needed to hear such high-pitched sounds; this ability is probably a legacy from their smaller canid ancestors. Foxes, which can hear the ultrasonic squeaks made by mice and other small rodents, use their high-frequency hearing to locate these animals when hunting. But wolves do not routinely seek out such small prey. ‘Silent’ dog whistles make use of the high frequency sounds that dogs can hear and we cannot, but they are something of a gimmick: whistles that produce at least some sound audible to human ears are much easier for us to control (how can you tell when a silent whistle is not working?). Dogs are, however, very skilled at distinguishing between quite similar sounds, probably using mainly high-frequency information. For instance, although research into how dogs discriminate between different types of barks is still in its infancy, there is little doubt that they can extract a great deal of detail from what they hear, as well as being able to detect very quiet sounds.
Dogs also have much more sensitive ears than ours, and as owners we should be attuned to this difference. Within their optimum frequency range, their hearing is approximately four times more sensitive than ours. This means that dogs’ hearing is probably damaged when they are subjected to the din encountered in some noisy kennels (which can be unpleasant enough even for us cloth-eared humans). Because of our own insensitivity to ultrasound, we are likely to be unaware of the discomfort that dogs must experience when subjected to noises that contain a lot of high-frequency sound, such as the banging of metal gates or the scrape of metal buckets on concrete floors.
In their sense of smell, dogs are miles ahead of us humans. And it is we humans who are unusually insensitive, not the other way around. Compared to other carnivores, dogs are about average. For example, grizzly bears have even more sensi
tive noses than dogs do, allowing them to find food underground even in the dead of winter. Nevertheless, dogs possess a unique combination of trainability and olfactory ability, one that we humans have made extensive use of throughout history – and indeed are finding new uses for almost every day.
It is hard to convey how sensitive dogs’ noses are without getting into some almost incomprehensibly large numbers. They can detect some odours, probably most, at concentrations in the parts per trillion. By comparison, humans generally detect odours in the range of parts per million to parts per billion, a sensitivity between 10 thousand and 100 thousand times lower than that of dogs. Dogs’ noses are as responsive as they are because they possess a very extensive olfactory epithelium. Although the area of this surface varies from breed to breed, the German shepherd’s, at 20–25 square inches (roughly the area of a CD cover, spread over a labyrinth of bony structures called turbinates) is typical – and over thirty times larger than ours. And between 220 million and 2 billion nerves, a hundred times more than in our own noses, link the epithelium to the brain. Why so many? Not only is the area of epithelium larger in the dog, but also the receptors are packed in much more densely on the dog’s epithelium than they are on ours. So that dogs can process all this information, their olfactory cortex – the part of the dog’s brain that analyses smells – is roughly forty times bigger than ours.3
Dogs can also pick up much more detailed information from odours, because they have a greater diversity of olfactory receptors than we do. So far, more than 800 functional olfactory receptor genes have been identified in the dog genome (along with a further 200 ‘pseudogenes’ that do not appear to make receptors, although they probably did at some time in the dog’s evolutionary past). Each gene codes for a slightly different receptor, each of which is sensitive to a slightly different shape in the odour molecules. Most odours trigger many of these receptors, and the brain then compares the relative strength of all the signals it receives in order to characterize each odour. Humans have a range of receptors similar to that of dogs, but fewer of each type. The implication is that everything that dogs can smell, we can too, but with less detail extracted. We also need a much higher concentration of most odours before we can detect anything at all. Humans can discriminate between thousands of different odours. Dogs’ much greater diversity of receptors suggests that they can detect a great many more.
In practice, the range of smells that dogs can detect seems almost limitless, judging by the proliferation of odour detection tasks they are asked to perform. Traditionally, mankind has exploited the dog’s nose in locating food, from the tracking of game to the detection of delicacies such as truffles. More recently, dogs’ keen sense of smell has been used to detect various types of cancer (melanomas as well as ovarian and bladder tumours) and impending epileptic seizures in humans. Dogs are able to smell pests such as the nematodes that can infest sheep as well as the bedbugs that can infest humans. They have even been put to use in conservation efforts: for instance, they are employed to sniff out illegal exports of sharks’ fins and sea cucumbers in the Galapagos. Scientists have likewise drawn on them to map populations of rare South American maned wolves and bush dogs (based on locating their faeces by their odour).
Dogs are much more capable than humans of discriminating between very similar scents. For example, they can distinguish between the odours of non-identical twins living together, and those of identical twins living apart. (However, they seem unable to reliably distinguish between the odours of identical twins living together.)4 In short, dogs can identify us using not only odour cues derived from the environment that we live in (for example, the food we eat and the fabric conditioner on our clothes) but also genetically based factors that contribute to our characteristic individual odours. Only when the genes are the same, and the environments also, do dogs begin to get confused. Dogs’ acuity at distinguishing particular human odours is now being used in several countries, including the Netherlands and Hungary, as a way of linking criminals to crime scenes.
Because we make so little use of our sense of smell, we have to exercise considerable imagination in order to understand how dogs experience this unfamiliar world. Odours do not behave in the same way as beams of light or sound waves; they are much less predictable than either. The rate at which they get into the air varies with temperature, humidity and the kind of surface that they are coming from. Moreover, the speed with which – and direction in which – odours travel is much more haphazard than is true of either light or sound. These factors do not matter much to us, and indeed rarely impinge upon our consciousness at all, because we use our visual sense to find our way around. Dogs, by contrast, have of necessity developed strategies to glean useful information from the odours that they rely on to locate objects of interest – whether those are scent-marks left by other dogs, potential food items, or odours that we have specifically trained them to find.
Finding interesting odours is not as simple or as instantaneous as gathering visual information. When we go somewhere new – for example, we enter a room we have not been in before – we look around and check our surroundings. Because light travels in predictable straight lines, it is immediately obvious if there are any parts of the room that we cannot see; we know without having to think about it that we cannot see into cupboards if the doors are closed, or behind screens or large pieces of furniture. Unfortunately for dogs, smells do not travel nearly so predictably as light. They spread very slowly of their own accord, by molecular diffusion, but the distances involved in this process are so minute (no more than a few inches) that they are relevant only to small insects like ants that live in the thin ‘boundary’ layer of still air close to flat surfaces.5 For an odour to travel any distance greater than this, it has to be transported by air movements, and these are very unpredictable and erratic.
To understand what a dog experiences, imagine opening a food cupboard and not being able to tell instantly whether something you are looking for is on a shelf inside, in a rack on the back of the door, or on the work-surface beneath. Try turning the lights out and locating a spice jar by its odour alone. We humans do retain some vestigial ability to navigate by smell, but it is a slow and cumbersome process. Even when we can smell something, tracing the source of the odour is rarely straightforward. Air movements are too unpredictable, especially indoors. That is why, if you watch dogs for a while, you will realize that they spend a lot of time and energy looking for visible indicators of likely places to find an interesting odour. How do they know where to sniff first? Presumably this is largely a matter of experience, though if they are leaving their own scents for others to find, they will either leave them somewhere obvious (on the proverbial lamp-post, say) or leave a visible indicator (such as the ‘tramlines’ that some dogs scratch in the soil, pointing to where they have just urinated).
If there are no visible cues, then dogs have to use their legs to work out where the smell is coming from. If there is not much air movement, they will run around sniffing, working out by trial and error where the smell is strongest. In situations where the odour is coming from a point source, this strategy is usually successful, sooner or later, but, if it is not, the dog can become very confused and frustrated. There is an apocryphal tale of a trained narcotics-detection dog that went crazy inside a container full of oriental furniture. There was a strong odour but he was completely unable to identify where the smell was coming from. As the story goes, it turned out that the whole consignment had been lacquered with cannabis resin – so the container itself was the source!
Dogs are particularly adept at following scent-trails, such as those left by other animals, or made deliberately by people, as in drag-hunting. Dogs will follow a trail by zigzagging to and fro. They find the edges of the ‘corridor’ of odour coming off the trail, and, if they lose it, will head back in the opposite direction, back across the invisible trail. As well as keeping to the trail, dogs have to make up their minds about which direction it is going in.
It is likely that wherever possible they use visual cues, such as the flattening of grass or undergrowth in the direction of travel. However, some dogs, though probably not all, seem able to follow a trail in the correct direction even if the visual cues are misleading. In one set of experiments, human subjects were persuaded to walk backwards across a grassy field, so that if the police dogs that then followed the tracks used heel-to-toe cues to determine direction, they would go the wrong way.6 Instead, they headed in the direction the persons had travelled. (It is possible, however, that they were picking up on the detail of which way the grass had been flattened, rather than simply the heel and toe impressions, even though the trails were by then an hour old.)
Bird’s-eye view of an odour plume. The average wind direction is shown from top to bottom, but changes in direction lasting a few seconds cause the plume to ‘snake’. Eddies (caused by irregularities on the ground) cause the plume to swirl and break up into pockets.
Dogs often face the problem of finding the source of an odour when there is no track along the ground. Outdoors, there is usually some kind of breeze to carry the odour to the dog, but breezes are not very predictable when it comes to carrying odour. You might think that odour travels in a straight line downwind from the source, but in fact it spreads out sideways as it travels downwind, resulting in a conical distribution with the source at the apex. Nevertheless, at any one instant, the distribution of odour would look, from above, more like a snake – one that is solid in some places and thin and wispy in others. The reason is that, as the wind blows over any surface, friction between the two causes eddies to develop. Some are several yards across, causing the snaking effect; others are smaller, causing the snake to spread out or bunch up. As a result, a dog that is standing still, even directly downwind of the smell’s source, will be outside the odour plume for a longer time than within it; conversely, a dog that is downwind but well away to one side will occasionally receive a burst of odour as the snake wiggles particularly violently.
In Defence of Dogs Page 28