3: Sniffing the Wind

Upton is a scent-on-the-air dog. He stands and sniffs in the breeze for many minutes, his head proudly up, eyes with a faraway look, cheeks puffing slightly with every exhale. I almost feel I can see his moist nose, lightly spritzed with mucus, catching odorous words from distant callers.

To understand how the dog can do what he does, you’ve got to follow his nose. What a dog is experiencing is formed of what he is smelling: as we see images in our minds, he glimpses scents; as we speak in words, he communicates in fumes.

Have you toured the dog nose? Ridden on a corkscrew of air into the dark vault, bumped along its curves, caught a breeze up to the chamber where a molecule will settle into the wetlands and begin to tickle the nerves to the brain?

I have—at least, near enough for my liking. I’ve been down the nose in a simulation of an odor molecule’s point of view as it is sniffed in. This unlikely video was created using airflow models generated by Dr. Brent Craven, who does work in computational fluid dynamics. Craven doesn’t have a stake in dog olfaction, as such. His research into how fluids and air move is primarily basic science, geared to understanding animal systems. Its application (and grant money) is often to reverse-engineer particularly good noses, by using airflow models for the creation of artificial noses for military purposes.

The nasal airway Craven and his team modeled is actually that of an eastern gray squirrel—a perfectly good smeller, but “much simpler” to model, he said. Imaginatively superimposed with MRI images of the innards of the more complicated canid nasal cavity, though, the video provides a sense of what a bumpy, hurly-burly, complex journey it is from the point of view of the sniffed odorant.

You’re riding on an odorant molecule. A miniature soap bubble comfortably bouncing and sailing on the briefest of wind currents, small and light and volatile. You’re in the vicinity of a dog nose, then suddenly, abruptly closer. The nostrils gape wider and approach. The odor bubble is sucked inside. Its speed is immense. Terrifying roller coasters could be modeled on the beginning of a long-nosed animal’s nasal cavity. For after a quick ascent, slowing as the path steepens, the odor’s arrival at the peak brings a heart-stopping view: only air ahead. A rush downward, curving this way and that, slowing for a second, then resuming. New projections pop from the walls as the odor races toward them, forcing you to bounce to your side, hit your head on the ceiling, and drop with a stomach-lifting thump. There are convoluted curves, perilous edges, and, all along, a gravity-defying force pushing you ever deeper. A tenth of a second later (slowed down two hundred times in the simulation), you are summarily deposited in a marsh well stuffed with moist grasses standing soldierlike, awaiting your arrival.

And this is all before you get to the brain.

NOSTRILY

Finn snortles up close to me to wake me. I reluctantly peek out with one eye. Its field of vision is entirely taken up with Finn’s right nostril, twisting this way and that, as we might do to make a silly expression. The result is that I twist my face into a smile and rise to greet him.

What has happened? At its essence, olfaction begins with the detection of as little as a few molecules of an odorant. The detector is the nose, the Hoover vacuuming in that trace amount. So let’s begin there.

If you have a dog near (and I hope you always do) bend down and take a good look at the dog’s nose. Get close, really close—one hopes, with a friendly dog who, at worst, will greet your examination with a peremptory lick of your own neb.

The hairy muzzle of the dog is what we usually look at—the snout, reaching back to the eyes and sloping on either side into jowls. Its length matters: while it isn’t the part of the nose responsible for the detecting, per se, nor the sniffing, it is the well-designed hallway down which the sniff will tumble.

Or we look at the moist, hairless tip of the snout, the so-called planum nasale or rhinarium. This is where the action begins. Dog nose tips are mesmerizing: as different as thumbprints, and as full of individual details. What most wet nose tips share is the pebbled topography of their surface, polygonal cells made visible in their opal darkness. This surface is wet so that more odors can be picked up and absorbed into the nose. Moreover, they have temperature gauges that can direct them toward a cooling breeze that might carry odors.

And their nostrilness! The nose tip carries two gaping nostrils—or nares—which will lead to the real nose an inch or six north, where special tissue lines the nasal canals. For those who find it hard to love a nostril, I say: You haven’t spent long enough looking at your dog’s. While to the uninformed they are just “holes in the front of the nose,” to those who study how dogs sniff, they are “articulating aerodynamic inlets,” their bulbous shape facilitating high airflow. Ringing the nostrils are the highly developed muscles in the alar folds. It is this anatomy that allows the nostrils to be an active part of the sniff, and that gives nostrils such a terrific range in shape from dog to dog. Many look like commas curled tight, or rounded into disks. Others are mere blunted openings, as if drawn hastily with a large-tipped pen. In some countries a noseprint is used as an identification: the inexpensive microchip. Ink the thin outer layers of the nose tip and press it against paper. A dog’s noseprint is the newborn’s footprint all parents will stare at for some clue as to their progeny’s future.

Dogs can use their nostrils separately and differentially: when sniffing something new and “nonaversive”—neutral or likeable odors—they begin with the right nostril, then switch to the left. By videotaping down the length of a cotton swab dabbed with an odorant while dogs sniffed, researchers found that lemon, food, and the secretions of a female dog elicited this right-then-left sniffing. When faced with the smells of adrenaline or the sweat of a kennel veterinarian (who swabbed his armpit for the sake of science), though, dogs sniffed only with the right nostril. It is thought that this nostril preference is due to the side of the brain involved. The right nostril connects to the right hemisphere of the brain (ipsilaterally, in contrast to the other senses, which cross sides)—which is more often associated with fear or aggressive behavior than the left, which analyzes familiar stimuli. If a dog is sniffing you right nostril only, he may be feeling suspicious.

Back to those muscles around the nares. Outside in a mild breeze, I watch as my dog Upton turns his nose—just his nose tip—to the right and left, bending in a way you don’t think a nose should or will bend. His nose gymnastics get the nostrils facing the breeze, picking up turbulent air, and snagging an odor. Then they open the aperture of the nostrils wide to maximize the amount of air that is pulled inside. An odor looks inside the business end of the telescope that is the nostril and closes its eyes for the ride.

SNIFFER

When I return to the house after an absence, I bend to Finnegan to let him sniff me, and to let me hear him sniff me. His sniff-snuffle is almost communicative—nosed phonemes in strings of merry sentences.

What we know about what happens when the odor faces its future up the dog nose comes largely from the work of mechanical engineer Dr. Gary Settles, one of Craven’s former professors. Settles, now emeritus at Penn State, has brought fluid dynamics—a field associated more closely with designing smooth-flying airplanes—to the study of noses. To my great delight, what Settles studies is the sniff. He and his team ask dogs to wear a specially designed muzzle, give them good things to smell, and measure the fluid dynamics of their sniffing.

Settles speaks of the dog’s nose as a classic “variable-geometry aerodynamic sampler.” The “sampler” (the nose) approaches a vapor cloud (air with smelly stuff in it) and transfers that cloud into its internal sensor chamber (the back of the nose, where the neurochemical magic happens).

The transfer happens through sniffing. As an invisible means of pulling in what are invisible odors, the sniff has long been downplayed. Further, while vision happens through light “hitting” the eye, in olfaction, odors do not just hit the nose; they sail right in. For that reason, smell feels intrusive. It is common for people to suggest that this is so because if one is alive and breathing, one can’t help but also be smelling: a gulp of air enters the nose with each breath, carrying, presumably, various odors. While it’s true that a nose-sized swig of air is bound to have smellable stuff in it, it’s not true that smelling happens by being alive and having a nose. It turns out that to smell, one needs to sniff. Not just breathe, not just sit with nostrils open. In the mid-nineteenth century, the start of the heyday of scientific discovery based on self-experimentation,I the physician Ernst Heinrich Weber, a founder of modern experimental psychology, first demonstrated the importance of the sniff. Weber lay down, had a solution of water and cologne poured into his nostrils, and waited, motionless. This is just the kind of experiment for which it is hard to recruit a lot of volunteers. What Weber found was that the normally ripe cologne was undetectable: he could not smell it when it was passively introduced to his nose.

Weber subjected himself to a noseful of cologne in the interest of discovering whether smells need to be delivered in vaporous form or can be detected in liquid. His was not the final word. A hundred years later, researchers renewed experimentation on the means of smelling: some repeated Weber’s supine nose-filling experiment; others presented strong smells before a person blowing through his lips like a trumpeter; and still others tried injecting odors intravenously to people during sleep apnea, between breaths. These studies had one common design element: in no case could the subjects sniff. The results? None of the bemused subjects perceived any odor. During typical breathing, only a very small amount of the air you inhale makes it up to the part of the nose holding the cells that will receive and manage the smell. One simply must energetically sniff air in.

So sniff, already! Sniffing is the inhalation part of a breath (you sniff in, but not out—that part’s more rightly a snort), often audible, requiring a modicum of effort. The variety of sniffs in the animal kingdom is certainly underplayed. Elephants sniff strongly, with trunk hovered over a scent or raised in the air for a “periscope” sniff. Gerbils sniff rapidly, noses twitching; tortoises, by contrast, have a slow-motion exploratory sniff in which they extend their necks, orient their heads down, and open their nostrils. A small puff of dust may appear on exhale. New Guinea singing dogs even “huff-sniff” when pursuing prey such as voles in a hole or under vegetation, blowing vigorously out the nose before sniffing vigorously in.

When canids first appeared, some tens of millions of years ago, their sniffing was enabled by the bellows that were the lungs, probably through a fairly straightforward nose. For fish, frogs, and reptiles, the snout is just a cavity allowing water and air to wash directly over the olfactory cells. This is still more complicated than the current invertebrates, many of which smell a lot, but not by way of sniffing smells in. Invertebrates have olfactory detectors outside their bodies, such as on antennae, so they must actually run into the odor source. Some—like lobster—send their sensory organ forward to sniff; others sniff simply by flying or wiggling their bodies and letting the air or water hit them.

Not for dogs, that crude method. Dogs not only have multiple sniffs, at different rates and for different purposes, but they have an ingenious method of exhaling that aids their smelling.

Settles discovered how dogs sniff by watching what he quaintly calls “canine nares airflows.” He recruited a handful of pets as well as detection dogs to come into the lab to be observed and recorded. There were golden retrievers, Airedales, Labs, German shepherds: all manner of noses, trained and untrained. The researchers put a variety of objects—highly desirable food, some inedible and novel scents, even a tiny bit of TNT and the sweet smell of marijuana—either right in front of the dogs or a distance away. Without being asked twice, the willing subjects sniffed. Detection dogs and pets sniffed alike, whether they were presented with pot or pet food. Both groups had two kinds of sniffs. When the smelly stuff is far away, unreachable, dogs do a “long sniff”—a highly directional, air-pulling sniff that lasts two seconds. In the long sniff, nostrils dilate and the alar fold opens; the mouth may be slightly open. Imagine the regal-looking large dog on a hill, his chest forward, his nose in the breeze. That’s the long sniffer. This sniff is often the culmination of a bout of sniffing, many weaker sniffs followed by a whopping one. In fact, one English pointer skilled at hunting birds was found to be able to maintain an extra long sniff. “Sir Satan”—the only of the tested dogs who was willing to suffer a sensor fastened into his nostril—maintained a continuous inhalation through his nose for forty seconds while running into the wind after a scent.

On the other hand, when the smell is nearby and on the ground, dogs sniff in bursts. First, they scan the surface. Have you ever noticed your dog nosing for a toy in a grassy area and seemingly not finding it, though she’s right on top of it? Her nose isn’t deficient: she is simply doing a survey of the area, as we scan a scene with our eyes. Typically a dog will walk her nose right up to the toy, actually pause over it, snort a little exhale, then continue on. To the vision-obsessed bipedal creatures nearby, it seems she’s missed it—rather stupidly—but she has not. She returns to the toy. She is simply assessing the concentration of all the smells in the vicinity to find the strongest source—something like a visual inspection of all the brunch buffet selections before you zoom in on the Belgian waffle (that you probably knew you wanted all along). Along the way, the dog does rapid “short sniffs,” from five to twelve every second, enough to make me hyperventilate just imagining it. This sniffing occurs at about the same rate as a dog’s panting—5.3 tongue flaps a second, on average—which is energy-efficient, and so might be their sniffing.

A hemisphere of air from four inches away is pulled up and in—this is called the nose “reach.” Given the chance, as every horrified dog owner knows, dogs will work to decrease that reach to a centimeter or full-on contact with the odor. Owners sometimes ask me, If their sense of smell is so good, can’t they smell it from here—i.e., at a polite, safe distance? But we mistake their noses for ours. They are not trying to detect it at all; they are trying to discern its contours, to take in all of its features, to make a measure of the smell.

Indeed, it is at a centimeter out that the smelly air is drawn in fastest. At this distance, the dog can get different odor samples from each nostril, which wind up giving the dog a bilateral vision that is a kind of “stereo” olfaction. Just as the images from our two eyes are constructed into a three-dimensional image of the world, the differences in the strength of the smell image in each nostril help the dog locate the source of the smell in space, whether it’s to his left or right, fore or aft.

Given this, asking why a dog is sticking his nose right in that dog’s rump is the equivalent of asking why you like to experience van Gogh’s The Starry Night from close enough to see each brushstroke—rather than peeking in the doorway from the next room.II

CANINE AIR JETS

But the most profound difference between the dog sniff and our own sniffing comes when dogs exhale. Our exhalations go straight out the entrance, out through the in door, pushily shoving any new air out of the way and preventing it from getting in. This can be a terrific relief when you want to purge a horrible odor from your nose, but it also sends lovely odors out almost as soon as they have arrived. When a dog exhales, he creates what Settles calls, charmingly, “expired turbulent canine nostril air jets.” Through high-speed videography of nostril and air motion, Settles found that dogs create tiny wind currents by exhaling not straight out, but out the side slits of their noses. This strategy minimizes the odor displaced—what Settles calls the “sample blow-off”—by the puff of air. The wings of the nostrils flare, a nose-plane ready for liftoff, and expired air leaves through a sneaky side exit. Not only does it not push the odor out of the way: the exhale creates a puff that lifts more smelly particles off the surface and a suction that hurries the next noseful of smell inside the snout. The exhaled nostril jets are little rotating funnel clouds that pull Dorothy, her house, and her little dog, too, right up into the nose.

images

Remember your dog’s short, thoughtful pause over the toy she is seeking? This is a pause of moment. She is sending her “expired air jets” right onto the source. More clouds of odorous particles come up from the toy and the ground. These jets are essentially increasing the reach of the nose, blowing and vacuuming in sync.

One scientist I spoke with likened this kind of sniffing behavior to the circular breathing a player of a wood or brass instrument might learn to do. It is the sniff without punctuation, allowing dogs to get a continuous read of the world—just in the way that we see the world without pauses while we blink.

Settles saw this jet-enabled sniffing by using specialized video called schlieren photography, which uses mirrors and slow-motion cameras to capture a picture of airflow. The photos make warmed air visible as blurred clouds emanating from the nose and mouth. In slow-motion schlieren videos, dogs’ snouts turn nearly prehensile, reaching and retreating to move the air; the muzzle seems to undulate, like a jellyfish pressing himself through the depths. But you can see some of this with the naked eye, too: simply watch your dog sniffing over a dusty plot of earth. After a particularly enthusiastic bout of sniffing some invisible thing, you can easily spot the puff of dirt and dust and odor pushed into the air and into the dog’s face with his own snorty exhale.

Could a sniff really be responsible for dogs’ keen sense of smell? Well, it is one of a number of key components. We know this because of evidence from when dogs lose their nose: when they are panting. Hot dogs can’t smell much. Dogs do not have sweat glands allowing them to release heat through pores across their skin. Dogs have only their waggy, pulsing tongues. They must pant—and panting, schlieren photos show, forces out such a wash of air that no smelly air can get to the nose. The panting dog must close his mouth to sniff well.

My Finnegan’s sniffing is audible, his own particular combination of huffing and grunting and spluttering. Outside, he is a nose-on-the-ground dog, trailing invisible stories in the grass. Give him a tree trunk that is popular with the local dogs, and he will serenade it with his rapid huff-splutter and his turbulent-jet exhales. At the other end of the leash, he will bring me to a stop with his nose set into an odor, his whole body cemented to the ground to allow the nose free smelling. Our Upton, the second dog, learned sniffing from Finn—learned that it was fine in this house to stand fast and observe the scents before moving on.

SNOUT

If one didn’t know which sense was most dominant in humans, it would take only a few minutes of observation to discover. Everything we want to perceive or examine, we parade before our eyes. Something to the side? We turn our heads (moving our ears away) so our eyes are nakedly upon it. Hear something overhead or underfoot? We don’t try to listen up or smell down: we turn our eyes. Our faces, while giving equal billing to the nose and mouth, have an array of protective devices around the eyes—eyebrows, top and lower eyelid lashes—and many of us use our noses simply as a perch for large spectacles to help us see better. Do you see? we ask, as a stand-in for understanding. Not Do you smell? Do you taste? Seeing is understanding, for us. When we meet each other, we greet each other with our eyes: to not look is considered rude, even abnormal. When walking, we start to turn our heads—and eyes—toward a corner seconds before we direct our legs around it.

So, too, with the dog’s nose. There is a fair bit about its simple position on the dog’s body that enables finely tuned sniffing. The snout is plainly protuberant. Not by coincidence, it is at the end of a head which, via a highly flexible neck, can reach the ground—where most odors lie. Dogs don’t spend a lot of time worrying about sniffing treetops: they sniff objects emanating from and landed upon the earth. As well as parts of other dogs’ anatomy that are, well, about nose-height.

Dog snouts are long for a good reason: evolution commits large amounts of anatomical real estate only to very useful causes. The cavelike, moist interior vaults under skin and fur are filled with air filters, humidifiers, and warming devices. When you’re sniffing dog rumps and decaying squirrel carcasses for a goodly portion of your life, you’d better have an excellent air filter. Inhaled air is cleaned and conditioned before it gets to the back of the nose. It must be dressed and ready for courtly presentation, as in the back it will meet kingly neurons that travel directly into the brain.

“So, you’ve been caught by a nostril,” Craven says. He’s walking me, riding that intrepid odor, through the interior snout. “The flow is pretty high in the vestibule,” the antechamber when the odor gets in the nostril. The dog is pulling in with his lungs through the nasal pharynx; the air is turbulent and chaotic—and then there is a fork in the nose. Inhaled air can go one of two ways: the breathing route or the sniffing route. If you go the respiratory route, you are heated up and humidified and proceed to the lungs. But if you’re sniffed to be smelled by a dog, you take a whole ’nother route, a high-speed ride toward the olfactory region. A current of air zips along a labyrinthine, tortuous path past a series of thin, curvy bones called turbinals; in cross section they resemble a big brain, convoluted and branched, folded into a small space. The turbinals, too, are part of the cleaning system—and, farther back, some are also covered with tissue useful for smelling. The turbinals create the roller-coastery pathways along which the odor rides on its sniffed journey.

Since air flows only one way across the long olfactory road up the nose (odors then are exhaled via the breathing route or are broken down by enzymes), dogs may be able to do something extra fancy with their noses: sort smells into groups as they fly in. Some odors are absorbed more readily than others, which means they will be grabbed by the sensory cells earlier in their route through the nose. For instance, researchers found that one component of the explosive TNT—DNT—is snagged more readily than other odors; this may partially account for sniffer dogs’ seeming ease at identifying it. It is more soluble, so dissolves earlier in the nose, than a molecule like amyl acetate (smells of bananas), which is itself more soluble than limonene (smells of lemon), which, like many other kinds of odors, will make it back to the olfactory recess before being absorbed into the mucus and finding a receptor to bind to. While we rarely ask dogs to find bananas for us, we could.III Given the layout of different sensory receptors throughout the nose, a dog can start to identify and discriminate smells in his nose before the brain even gets involved.

Toward the back of the channel the odor suddenly slows. The turbinal bones here are lined with olfactory epithelium—brown tissue that holds the sensory cells, which will snag odors out of the air and is where the magic conversion from “odor” to “odor you smell” begins to happen. The cells are coated in mucus.

Then, “At some point, you will be deposited or absorbed by the mucus that’s lining these airways, and you’ll slowly diffuse,” Craven warns. It’s not as grim as it sounds. A thin lining, about ten microns thick, forms the transition from the outside air to the internal neurons. The odor you are riding will migrate through the mucus in a tenth of a second, about the time it takes for a sniff to travel from the nostril to the back of the nose. But for now you can relax.

RECESS FOR ODORS

You’ve arrived at the cul-de-sac of the snout, the olfactory recess. This is the part farthest back in the nose, at a point a half inch into the skull, just between the eyes. In the olfactory recess odors can linger, trying to find a sensory cell to nuzzle up with for many rounds of inhaling and exhaling. Dogs get a chance to really ruminate on the smells they’ve sniffed in before the air is scooted out.

The recess—as well as some of the bones along the way—is lined with the aforementioned epithelial tissue. When newspaper articles say that a dog’s sense of smell is ten thousand times better than a human’s, or one million times better, or whatever-exponent-of-ten times better, one of the most cited bits of anatomical proof is the amount of olfactory epithelium—the expanse of nose with cells specialized for smelling smells. Though the comparative numbers are suspect (and quite variable for different odors), dogs are massively better endowed than humans in olfactory sensory cells. If his olfactory epithelium were spread out along the outer surface of the dog’s body, it would completely cover it. In humans, ours would about cover a mole on our left shoulder.

The epithelium here is covered with a dense mat of cilia, little hairlike branches that extend from the sensory neurons. A few dozen cilia sprout from each nerve, and each is coated with dozens of proteins called olfactory receptor cells. Receptors do what they sound like: they receive odors. To do so, they stick, untroubled, right out there in the mucosal environment of the nose, perfectly designed to snag odorant molecules coming by on a sniff.

Dogs cram more smellability into their snouts at every level: they have many cilia on each neuron and more receptors on each cilia than humans do. Indeed, every dog has hundreds of millions more cells devoted to detecting smelly stuff than humans do. Dogs have from two hundred million to one billion receptor cells, depending on the breed, compared to the six million in our noses. In the dog’s case, more nose mass also enables more kinds of receptors—over 800—that themselves can encode more information about the odors.

This number—eight hundred and change—gives researchers pause. The eye, conveyer of the brilliance of a sunset dancing off the clouds after a thunderstorm, uses just three receptors to draw this colorful scene in our heads. With eight hundred more receptors, the possibilities for the odor landscape are staggering. The number of odors the dog could detect could theoretically be “billions,” Dr. Stuart Firestein, a neuroscientist at Columbia University who studies olfaction, has written. But “in fact, the question is probably not relevant, just as it makes little sense to ask how many colours or hues we can see.”

Even when odors land in their receptors, they are still undercover. The nose does not know what they are. There is no “cheese” receptor, activated by the Stiltons and cheddars my dogs nose out on our kitchen counter. Despite the dog’s alacrity in locating the corpse of an unlucky squirrel in the park, there is no “dead squirrel” receptor. Simply, each odor activates many receptors; there is not one receptor for each odor.

Although the means of reception has not yet been conclusively determined, the most popular theory of how it happens uses a lock and key metaphor. In this model, receptors are the locks, of various shapes and lengths, and the various molecules that make up odors are the keys. A related theory suggests that the reception of odors is less specific than a key in a lock; instead, it’s more like a key in a pocket, in which many differently shaped keys can bind to a receptor and prompt it to fire. Apt for dogs, smeller of pockets.

When biologists Drs. Linda Buck and Richard Axel won the Nobel Prize for their work in olfaction, it was for discovering the genes that code for these receptors. Amazingly, olfactory genes are wildly overrepresented in the mammalian genome. Dogs have around eleven hundred olfactory receptor genes, some eight hundred of which are operative.IV Keep in mind that your dog’s genome—the blueprint for making his entire body, from his charismatic curled tail to dark expressive eyes—is just over 19,000 genes. Nearly 5 percent of the genome is committed to making smell receptors alone—plenty of kinds of locks with which to smell the world’s keys (or pockets).

Breeds vary in their smelling abilities, and it may be because they have varying numbers of operative olfactory genes: boxers (short-nosed, compressing that turbinate-filled snout) have slightly fewer functional genes than poodles (long-nosed and respectable smellers). Though research has just begun on the topic, there is some evidence that specific genes might even be linked to detecting specific odors. In one study, a tiny substitution on a particular gene was seen in dogs who did less well on explosives-detection work.

If genetic differences lead to differences in detection ability, one may well ask, does this mean that certain breeds are genetically superior sniffers? Insofar as some undetermined subset of a dog’s genes may lead him to notice a smell that another dog does not, yes. Whether the genetic difference in smell abilities is part of the genetic difference in breeds is another question—an unanswered one, for now.

THE SMELL HITS THE FAN

All that work to get to the receptor cell—but the odor is not done yet. The smell of dead squirrel—or any other substance—is discovered only when components of its odor, having snuggled into receptor sites, cause the neurons to fire—changing voltage—and sending an action potential down its length, leaving the nose and entering the nose brain: the olfactory bulb. Tens of millions of neurons converge into a few thousand bundles and sneak through small openings in the bone into the brain. Long ago, it was thought that brains did the smelling, and the nose was just the conduit. Even in the twentieth century, with the rise of brain/computer analogies, it was not uncommon to hear the nose be called simply the fan of the massive supercomputer that is the brain. Now we know it’s a bit more electrical. Santiago Ramón y Cajal, an early and influential anatomist, mapped the route from the nose to the brain at the end of the nineteenth century, observing that nerves (carrying word of odors smelled) enter the brain, not the odors themselves.

The olfactory bulb sits right behind the back of the nose, jammed in under the frontal lobe. The nose is the quickest route to the brain: one neuron goes from the warmed, dinner-smelling, light-spritzing-of-dog-hair-wafting air in your living room to, on the other side, the highly conditioned environment of the brain. Before the olfactory bulb, all a neuron “knows” is that it’s firing; in the bulb, axons from thousands of neurons from the same kind of receptors all converge on a single target site and flood it with activity. What this seems to enable is the cobbling together of a smell sensation. Just as an odor molecule is grabbed and deconstructed by the receptors, causing many to fire, it is reconstructed in the topographical layers of the olfactory bulbs. Little traces of evidence from many cells are translated into the fetid, cadaveric smell sensation that is your dog’s discovery.

One might expect the dog’s olfactory bulb to be humongous. It’s not. But it is 2 percent of his entire brain, two pennies to the silver dollar of the brain. (In humans it is vanishingly small: less than a thirtieth of a penny.) And this matters: the olfactory bulb translates those receptor cells firing into something like an experience of an odor.

From the olfactory bulb the odor information is zipped through the dog’s brain on a massive and near-instantaneous search for recognition, how to feel about the smell, what memory it might awaken, and what behavior it might prompt. The bulb connects straight to the olfactory cortex for some of these decisions, as well as directly with the subcortical limbic system, which adds the emotional tone—fear, excitement—to the odor.

Research on the brain’s response to odors has always had one common element: present the subject with something truly stinky. Measure the result. An early researcher studying the brain’s response was physiologist Edgar Adrian, who exposed the humble hedgehog to various stinks, including the odor of decayed worms. “Water in which an earthworm had been allowed to rot” prompted a big reaction in the hedgehog.

More contemporary researchers have presented rabbits with store-bought sharp cheddar, lured tsetse flies with a combination of oxen and buffalo urine, exposed terrified rats to anal gland secretions from weasels and red foxes (rat predators), and, most recently, stuck the smell of their owners’ armpits under the noses of dogs.

To back up. Neuroscientists can be fond of trying to answer the big questions about the mind with machines. MRI machines, when you are in one or looking at images generated from one, can seem superficially like mind reading. In “functional” MRI, a subject lies on a platform surrounded by walls housing a very powerful magnet. Through disturbance of the magnetic field, images of the blood flow in the brain (indicating neural activity) can be captured. Lie down, think of your grandmother, and the areas of the brain where your memory of her crooked smile and glasses, the smell of talcum powder, and the melancholy memory of the toys she made for you as a child reside will light up on the computer screen.

fMRI will never answer the questions of what it is like to have those memories or to revisit her smell. Images of brain location do not explain how the odor of talcum prompts me to remember my grandma sitting in her special chair in her dark, stuffed living room. Instead, the device lets us view the activity from afar: watching the Perseid meteor shower without answering the questions of the universe. And so a few research programs have begun looking at dog brains in MRI machines. This is itself a feat of training and patience—though quite doable for both species—as the dogs need to be very much awake in the machine, and yet stay perfectly still. One early study looked at what areas of the dogs’ brains were triggered by their owners’ smells.

By “owner’s smell” I mean “smell of the person’s armpit,” caught on a gauze pad and waved in front of the dog prostrate in the MRI magnet. They found an area called the caudate nucleus dancing with excitement at the armpit gauze. Researchers had looked there because it’s easy to visualize in the MRI, and it is associated with reward.

I suggest that another approach to discovering where the olfactory information goes in the brain is to watch your own dog. If you live with a dog, you know what happens after dead squirrel is identified by the brain. No dog waits and considers the wisdom of his next move before doing it.

The dog rolls in it.

This jibes with a theory Stuart Firestein advocates. “My craziest idea, you want to hear it?” I do want to hear it. “I think the actual percept of an olfactory object”—the sense of experiencing a smell—“may not occur until you hit something very close to or actually in the motor cortex,” Firestein tells me. “Because most of what we do in olfaction is very tied to decision-making: do you put this in your mouth, do you flee from it, do you fuck it, do you whatever-you-want-to-do-with it?”

Indeed, some signals from the olfactory bulb do go straight to the motor cortex, a part of the brain that controls movement. It directs the muscles in the dog’s legs that cause them to buckle gracefully, the precise angle of the head to squish perfectly into the smelly stuff, the intensity of supine squirm necessary to distribute the scent down his back. Note that when the dog gets up from throwing his body on and thrashing about in that squirrel, the first thing he wants to do again is get another whiff.

VNO

He licks. He licks my knee, my face, my ear. He licks the air, tongue reaching for me, as he approaches. It feels like affection, and I smile. But he also licks the corner of the building, the very smelly-looking patch of grass, and—sigh—the cat’s bottom.

So much nose! But in dogs, and in many mammals, the olfactory system is bigger than just the nose. Dogs have a kind of “second nose,” right under the bone separating the nostrils and above the roof of the mouth. Two scrolls of cartilage house what’s called the vomeronasal organ—snappily abbreviated to VNO by researchers wanting to save themselves four syllables—which is as much a part of smelling as the nose is. Despite its evocative name, vomer describes the shape of the organ: ploughshare-like, the part of the plow that does the cutting. Secreted under the nose, the organ cannot be reached by simply sniffing; instead, odorants must dissolve into the tissue and be sucked inside. The pumping mechanism is prompted either by touching the molecule directly or through making a horrifyingly silly face called flehmen. Should you see a horse curl back his upper lip, seeming to scowl and shuddering a bit, you are witnessing classic flehmen: it pulls the odors back to the nasal tissue to be absorbed. To flehmen, the pig opens his mouth wide; the cat holds her mouth a small amount open, giving her a disconcertingly muddled appearance. A snake’s forked tongue flicks to pick up odors and send them to each side of its VNO.

Most dogs do not do a classic flehmen lip-curl, but they have their own methods. At times, after sniffing, a dog may wrinkle his nose in an apparent grimace and chatter his teeth—that’s a canine version of flehmen. Even better, they lick. The dog’s extravagant, long tongue is good for reaching the bits at the bottom of the peanut-butter jar and cleaning your legs of post-exercise sweat, yes, but it is also a perfect mechanism for bringing odors to the VNO for investigation. Lick ground, lick nose, smell.

What the VNO enables dogs to detect is a kind of molecule that the ordinary olfactory route often cannot, such as pheromones. A pheromone was originally defined as a signal sent between two members of the same species—which causes the receiver of that signal to behave or develop in a very specific way. Androstenone, produced by a boar, causes a sow to assume, rather robotically, a mating posture; bombykol, given off by the female silk moth, wafts onto the antennae of the male silk moth and causes him to seek her out. Pheromones are used by an incredible diversity of organisms, from lobsters and rabbits to ants and bacteria.

What makes these pheromones detectable by the VNO is that they are typically water-soluble chemicals, nonvolatile, low-molecular weight molecules. So are lots of other molecules, such as hormones and “signature mixes” that may hold information about an animal’s identity or the particular family or pack it belongs to. The VNO receptors are tuned to be highly specific and sensitive, unlike the more broadly receptive olfactory receptors in the nose. The rump-to-face-to-rump sniffing dance between your dog and another dog is a chemical communication of their sex, readiness to mate, their health—and also who each of them is. Each animal’s urine and saliva may hold the same.

SNEEZER

Apart from the snout proper, the featured player on the face, the dog’s sense of smell commandeers other dog features and behaviors for its olfactory ambitions. Not for nothing are bloodhounds so well ear-endowed. James Thurber’s famous sketches of the breed have as much ear as head. And ears help make members of the breed such outstanding smellers. With his nose pointed down, the bloodhound’s long ears—over thirteen inches, in extraordinary cases—sweep up odors from the ground for better smelling. They are dual fans attached to the face to direct possible smells to the smeller. Even the drool that hangs from a bloodhound’s mouth might help to bring odors upward to be absorbed by the vomeronasal organ for consideration.

I discovered two other ordinary behaviors involved in smelling through simply watching dogs. Many of my research programs with the Dog Cognition Lab are rooted in observations of dog behavior in natural environments—for instance, in public parks, around people and other dogs. I have spent untold hours videotaping dogs playing together, then coding their behavior by reviewing the tapes one-thirtieth of a second at a time. I code the video in this super-slow-motion way because it enables me to see things we miss in real time. But there is another benefit: watching video at that pace defamiliarizes the dog. While it is terrific fun to watch dogs playing (believe me, it was hard to convince a dissertation committee that this might be actual work), our natural human tendency is to see the play but not see what is really going on. To see, but not see.

When we see dogs, we are so immediately and confidently sure of what they are doing—Oh, they’re friends; look, he wants to join the play; she is shy—that we preempt actual examination of what they are doing.

It was while out watching dogs in this way that I made a surprising observation: dogs wag their tails a lot while meeting someone new or greeting a dog they know.

Wait, you say: that is not new! Of course. We know what tail-wagging is for. A loose, jangly, high tail-wag is a sign of friendliness. A fast, low tail-wag indicates anxiety. This is true—but simply knowing this makes it harder to see what else tail-wagging is doing. It is spreading scent. Whether intentionally or not, when a dog wags her tail, all the very fascinating odors (to a dog) from her anal sacs spread in a bloom around her body. She is not only telling other dogs how she’s feeling; she’s scenting who she is.

There is precedent for the use of the tail-wag as scent-spreader in other animals. The hippo madly fans his tiny tail while urinating and defecating, better to atomize that scent. Some rodents and squirrels wave their tails when a possible mate is nearby. Woodland caribou, whose tail is covered with scent glands producing a musky odor, use it as a “scent brush” to disseminate an alarm odor. So though a dog wagging his tail looks simply like a dog wagging his tail happily, he is also saying Come smell the wafty odor that is me.

Similarly, walking with my dog Finnegan one day, I noticed another usual but unusual behavior. As we passed a tall, prancing black poodle being walked on a tight leash across the street, our dogs looked at (smelled at) each other. Suddenly, the poodle did a full body shake—while still walking. I could very nearly see the plumes of scent from her curly hair hurtling toward Finn’s nose in a fume cloud, and he was fixated. Could a shake be deliberate? I wondered. It might be the canine equivalent of flirtingly running my fingers through my hair as I talk with a suitor, perfuming the air around me with the scent of some slightly overpriced shampoo I use.

Trying to see through the familiar dog behavior also allowed me to learn about the sneeze. Sneezes are sneezes, to be sure: reflexive ways to clear the nose of something ticklish or alien. Dogs sneeze only through their noses, as opposed to the human mouth-nose sneeze. The nose-only route marks an interesting secondary use of sneezes: to clear the nose of an undesired smell. In this case, I believe dogs use sneezes intentionally to end a session of smelling one odor and to ready themselves for a different one. Watch your dog sniff a strong, recently laid smell at a corner, then see if she doesn’t sneeze before moving on down the street. Relatedly, trainers of tracking dogs have noted that some dogs clear their noses by lifting their heads up from the ground, taking in a bit of less scented air. It is not a pause or a miscue; it is an essential part of his tracking.

•  •  •

In the end, it is not the size of the olfactory bulb itself that makes the dog’s nose so keen. It is not the numbers of receptors per se, or the way that the dog sniffs. It is not just the length of his nose. It is all of it. It is sniffing in the way he does, through the nose he has, into the many receptors he houses, to the brain that has evolved. The result is astonishing.

You have surely seen the sleeping dog who, eyelids fluttering, toes dancing, barking muffled through a closed mouth, seems to be dreaming. Dogs do have REM sleep—the time when we most often dream—so it is likely that they are also dreaming: chasing, with those dancing toes; announcing, with that muffled bark. Next time you see your dreaming pup, look for his active nostrils. For a creature of the nose is almost certainly smelling in sleep—following the odor of his friend down the street or of dinner newly plopped in a bowl, or investigating a curious smell arriving under the door.

Go and look at your dog’s nose again. I love to look at my dogs in their eyes—a shared glance is full of understanding, an agreement that we belong to each other. That dogs look at us is what bootstrapped an ancient proto-dog, lurking around the edges of early villages, into the dog who sits on your lap and wags at the sight of you.

But now I look at my dog’s nose, too—the whole length of it, but especially its wet tip—and my heart also skips. What moves me is the depth of what that nose knows about the world.

•  •  •

Talk to human psychophysicists and neuroscientists, and they’ll say, You know, it’s a myth that dogs smell so much better than humans do. And you can look at them and ask them what they smelled today. The answers vary from “nothing” to the poetic: “grassy meadows, prairie grass,” smell scientist and author Dr. Avery Gilbert reports to me, of his new home in Colorado. Pretty good. Now consider asking a dog. If a dog could talk, his answer would be an epic poem recited over several hours. In fact, humans have fine noses. But most of us simply don’t bother to smell. I mulled this over as I sniffed Finnegan’s fur one night (smell: fresh running river, water burbling over river rocks). Hmm. Maybe I’d bother to: to try to make my nose a worthy companion to my dogs’.


I. Practitioners included Freud, who famously experimented widely with cocaine; plenty of others who ate or drank questionable (including radioactive) materials; and researchers who self-catheterized or gave themselves viruses or unfinished vaccines.

II. Yes, I am equating the dog rump with The Starry Night. Dogs have perfectly artistic rumps.

III. And, in fact, there are now many detection dogs trained to find bananas—as well as any other agricultural contraband brought across the border illegally.

IV. The rest are “pseudogenes”: genes that have a mutation such that they no longer result in the development of a receptor—that is, they no longer make the thing that they are supposed to. For dogs, 20 to 25 percent of their olfactory receptor genes are pseudogenes. For humans, it’s more than 50 percent.