What the Nose Knows: The Science of Scent in Everyday Life (10 page)

How good a nose does a government agent need to claim probable cause for a drug search? The Ohio court relied on the fact that the arresting officer was trained and experienced in identifying the smell of marijuana smoke. Other jurisdictions aren’t so fussy. The degree of nasal prowess claimed by police officers can, at times, beggar belief. In New Jersey, for example, a driver was pulled over for a traffic infraction. The police officer claimed to smell fresh, unburned marijuana through the open driver’s window. A search revealed a brick of Mexican pot wrapped in a plastic garbage bag in the trunk, where it had been placed after a drug buy twenty minutes earlier. In California, police searched a house where they suspected pot was being grown. They didn’t obtain a warrant because they claimed they could smell marijuana plants—from several hundred yards away in the hot air and diesel exhaust venting from the chimney.

These feats of nasal detection are all the more remarkable given the level of training most police officers receive. According to Jim Woodford, who serves as an expert witness in criminal trials, officers often learn drug smells by sniffing the real thing in the evidence room. Formal training is rudimentary. “Somebody comes in with a suitcase of stuff, everybody goes by and takes a sniff. That’s the training,” he says. The problem with this informal approach is that marijuana aroma is extremely variable, something potheads are well aware of. (Just ask the reviewer of the Beck concert in Costa Mesa…)

Of course, police officers become familiar with drug smells while busting dealers and users. They cite this on-the-job experience when defending their skills in the courtroom. They testify that “I’ve been on so many busts, and I recognize it. Over the years I’ve learned it.” Woodford says, “That’s sufficient to be deemed an expert by the court.” He says it is rare for the defendant in a drug case to challenge the officer’s smell ability via a smell test or medical exam.

Just how detectable is the smell of pot under circumstances such as these? Richard Doty and colleagues conducted some forensic sniff tests to find out, using experimental conditions modeled on the New Jersey and California cases. They found that untrained people can easily distinguish a Hefty bag containing 2.5 kilograms of Mexican pot from one holding an equal weight of shredded newspaper. But when the samples were placed in a car trunk, the panelists could not detect the smell from the driver’s window. Likewise, panelists could reliably identify mature female cannabis plants at close range by scent alone, and could distinguish immature pot plants from tomato plants by smell. But when the smell of marijuana plants was mixed with exhaust from a diesel generator, the panelists couldn’t detect it.

When it comes to detecting drunk drivers, sniff-based forensics are on even shakier scientific ground. A study by the National Highway Transportation Safety Administration found large variability in the ability of police officers to smell alcohol on a person’s breath. As a group, cops picked up the scent consistently only when the drinker had a very high blood alcohol level (the detection rate was 61 percent for BACs between 0.10 and 0.15 percent). In the most rigorous study on the topic, all variables except odor were eliminated: test subjects were hidden behind a screen and breathed at the officers through a tube. The police participants were all highly experienced and trained as Drug Recognition Experts. Even so, test performance was highly variable across officers. As a group, they detected breath alcohol 85 percent of the time at BACs of 0.08 percent or more, but caught it only two-thirds of the time at lower levels. An officer’s ability to estimate the intensity of breath alcohol odor was no better than chance.

T
HAT POLICE OFFICERS,
like everyone else, show a wide range of olfactory ability comes as no surprise to smell scientists. That their abilities should be granted special consideration by judges and juries is another matter. Doty and his colleagues argue that skepticism is in order when marijuana is said to be “in plain smell.” Sensory claims by police are least substantiated when it comes to fresh, unburned marijuana. Yet this is just the circumstance where courts have given greatest credence to a police officer’s nose—no corroborating evidence is needed. Doty’s study has already been cited by the defense in a drug case in federal court. (A police officer with no training in pot aroma claimed to smell immature marijuana plants in an unvented grow house from a long distance away.) Can trained police officers outsniff civilians? Probably. But according to Doty, this has yet to be scientifically documented. Helen Keller would expect better from the Federal Bureau of Aromatic Specialists.

CHAPTER 4

The Art of the Sniff

The smoke of my own breath;

Echoes, ripples, buzz’d whispers, love-root, silk-thread, crotch and vine;

My respiration and inspiration, the beating of my heart, the passing of blood and air through my lungs;

The sniff of green leaves and dry leaves, and of the shore, and dark-color’d sea-rocks, and of hay in the barn…

—W
ALT
W
HITMAN,
Leaves of Grass

S
OME SMELLS ARE MORE SUBTLE THAN OTHERS.
T
HEY
float up the nose on the tidal rhythms of normal breathing and may not reach conscious awareness until minutes later. When we want to pay attention to an odor, we don’t wait for the next lungful of air—we capture it with a sniff. Sniffing is an odd behavior—it has no analog in vision or hearing. (Dogs, mice, and deer can rotate their external ears to focus on sounds; we can’t.) Sniffing is ignored by students of “body language.” It can be done covertly, and in polite company it usually is; sniffing is considered rude, and audible sniffing is downright vulgar. It takes an uninhibited, bumptious soul like Walt Whitman to draw attention to it, much less revel in it. But there is no getting around it; sniffing is essential. Whether one is tracking down a dead mouse in the basement or savoring a newly opened bag of Doritos, the sniff is the prelude to a smell.

The purpose of a sniff is to get scent molecules to the place where we can smell them. The question that took philosophers and scientists thousands of years to answer was, Where exactly does smelling happen? Some ancient Greek philosophers argued that it took place in the nose, but the sievelike appearance of the cribriform plate—a bone at the base of the skull just above the nasal passages—led others to speculate that odor particles made their way directly to the brain through these tiny holes. In this view, the nose is a merely a tube and the brain is the sensory organ. The ancient nose-versus-brain debate wasn’t settled until 1862, when a German anatomist discovered the olfactory nerve cells in a cleft high in the nasal passage. Smell—at least the first physiological contact with odor molecules—clearly happens in the nose. The holes in the cribriform plate are there to allow nerve fibers from the sensory cells to reach the brain.

Because the olfactory cells were tucked away in a narrow olfactory cleft, they did not appear to be exposed to the main flow of air through the nose. Researchers were soon asking how much of air entering the nostrils actually made it to the olfactory nerve endings. Early experiments were ingenious and also a bit macabre. In one study, for example, the head of a cadaver was cut in half and tiny squares of litmus paper were placed throughout the nasal passages. The head was reassembled and ammonia vapor pumped through the nostrils and out the trachea. Color changes in the papers showed that very little ammonia-laden air made it to the sensory cells; most passed through the lower passages. A second, more grotesque experiment anticipated the slice-and-shock art of Damien Hirst by a century. A split cadaver head was pressed against a glass plate and smoke was blown into the nostril. Observers could see the currents and eddies as the smoky air flowed through the complex folds of the nasal chamber. The smoke patterns, like the ammonia vapor, showed that only a fraction of the incoming air made it to the receptors.

Today, sophisticated computer models can simulate nasal airflow. Researchers can see where the flow is laminar (smooth) and where it is turbulent. They can calculate how many scent molecules are deposited onto the sensory surface as air is drawn across it. For all the high-tech apparatus and numerical precision, the modelers reach the same conclusion as their head-splitting predecessors: only about 10 percent of inhaled air blows across the nerve endings in the olfactory cleft.

T
HE SNIFF
—a short inhalation with a high rate of airflow—is an essential step in odor detection. By forcing more air past the olfactory cleft, we take a bigger sample of the external smellscape. So how did it come to be dismissed and even suppressed by serious scientists? This is a strange tale. The first scientist to pay much attention to sniffing was also the one who tried to eliminate it from smell experiments. In 1935, Charles A. Elsberg was a highly regarded neurological surgeon in New York with a flair for invention—he designed surgical instruments and had performed the first successful removal of a herniated spinal disk. Elsberg’s flair for promotion was even bigger. He had cofounded the Neurological Institute of New York, set up the country’s first Neurosurgery Service there, and later cofounded the Society of Neurological Surgeons. At the age of sixty-four, it occurred to Elsberg that brain tumors, by exerting pressure on the olfactory areas at the base of the brain, might lead to impaired odor perception. If he could measure odor sensitivity, he might be able to identify patients with brain tumors. Accordingly, he came up with a method that involved a bottle, a cork, a syringe, and some rubber tubing. The patient would hold his breath and Elsberg would inject odorized air into his nostril. Acuity was measured by how big a blast of air was needed for the patient to detect a smell. Elsberg found that a normal person needed six to nine cubic centimeters’ worth. Elsberg’s system was coldly efficient; it not only eliminated sniffing, it eliminated breathing.

Elsberg touted his method as a major breakthrough: the first scientifically objective measurement of odor sensitivity. He either didn’t know of, or didn’t care to acknowledge, the olfactometer invented thirty years earlier by Hendrik Zwaardemaker. Every sensory psychologist in America was familiar with Zwaardemaker’s device, and most had one in the laboratory. It consisted of a glass sampling tube, curved at one end to fit into a nostril. A wider tube, containing an inner layer of scented material, fit snugly over the sampling tube. The farther the wide tube was pulled back, trombone-like, off the end of the sampling tube, the more scented surface was exposed. Sensitivity was measured as the length, in centimeters, that the scent tube had to be withdrawn in order to create a detectable level of odor. Zwaardemaker’s device, of which several versions were available, was reliable enough to explore the basic phenomena of odor perception and was used in laboratory demonstrations in colleges across the country. Nevertheless, Elsberg’s results were soon written up in
Time
magazine and on the front page of the
New York Times.
In the latter, the headline read, “Brain Tumors Detected by Scent with Device Keener Than the X-ray; Neurologists Hail Dr. C. A. Elsberg’s Discovery as Epochal—Based on Accurate Measurement of Sense of Smell, Which Was Viewed as Impossible Heretofore.” According to the credulous report in the
Times,
“Dr. Elsberg succeeded for the first time in measuring what had hitherto been considered universally as unmeasurable. He established a definite ‘scent yardstick.’”

Having nine cubic centimeters of air rammed up one’s nose is no barrel of laughs. However, blast injection proved to be a popular technique: most scientists prefer tight experimental control, even when precision comes at the cost of realism. Eventually researchers grew skeptical about the Elsberg method. They found that blast volume mattered less than blast force—this undercut the use of volume as a measure of smell ability. Even more troublesome, blast force was irregular—it depended on how abruptly the experimenter released the pinchcock on the rubber tube. The enthusiasm for nostril-blasting ended in 1953 when a psychology professor at UCLA compared odor sensitivity measured by Elsberg’s method and by natural sniffing. Blasting produced unreliable data, while natural sniffing produced very reliable data. The results blew Elsberg out of the water. Blast injection was not the scent yardstick he claimed it was. As the syringes and hoses were packed away for good, another psychologist ruefully wondered whether “we might be better off today if Elsberg had never publicized his creation.”

Mr. Natural: Keep on Sniffing

The physical characteristics of a sniff are smell dependent. Confronted with a weak scent, we take larger and longer sniffs, and more of them. We take smaller, shorter, and fewer sniffs to a strong odor. Considering how essential sniffing is to smelling, one might think this behavior would be studied by many scientists. Yet the bulk of what we know about sniffing is largely thanks to the work of one person, the Australian psychologist David Laing. He pioneered the natural history of the sniff.

In a series of elaborate studies beginning in 1982, Laing established how the dynamics of sniffing relate to smell. He controlled what people smelled with an air-dilution olfactometer, a device that generated a stream of air with precisely controlled odor levels. He measured how they sniffed by means of an oxygen mask with a tiny airflow probe concealed in it.

Laing found that natural sniffing took place in an episode of three and a half sniffs on average; some people used fewer, some many more. A person’s sniff episodes have a characteristic pattern that is stable across different odors and tasks. Sniff patterns were so stable and individually distinctive that Laing found he could identify a person by airflow data alone. He went so far as to liken sniff patterns to fingerprints.

At the time of Laing’s work, I was beginning my first experiments on human odor perception at the Monell Chemical Senses Center in Philadelphia. My odor sources were plastic squeeze bottles with fliptop caps. I would sit behind a screen and hand one bottle at a time to my test subject, who would squeeze, sniff, and rate the odor. As I listened to the wheezing of the bottles, I realized each person had a typical sniffing style. I soon developed a private taxonomy of sniffers. There were the Delicates, who took tiny, barely audible sniffs. There were the Honkers—people who squeezed the hell out of the bottle and inhaled so forcefully I thought they might hurt themselves. I also observed different psychological profiles. There were Decisives—people who sniffed and promptly announced their rating—and there were the Agonizers, who sniffed and resniffed and sniffed again before summoning up a rating. Every combination of behavior and decision-making style turned up in my lab: Delicate sniffers who were very decisive, Honkers who were Agonizers, and so on. These patterns were so consistent that after two or three squeeze bottles I could predict how long the entire test would take. A diverse range of local oddballs answered our recruiting ads. Once, in the middle of a test, my research assistant handed a sample of patchouli around the screen. There was some squeezing and sniffing, followed by a long silence. Finally she looked around to find that her subject had poured the sample into his hand and was massaging it into his beard. He said he liked how it smelled.

Intuitively, it seems the more one sniffs, the better one smells. Like dogs at a fire hydrant, multisniffers must be extracting every last bit of information from a smell. But are they? David Laing systematically controlled sniffing to see how it affected a person’s ability to detect and describe a smell. Sometimes he allowed his subjects to sniff with their natural pattern; other times he told them exactly how many sniffs to take, how long to wait between sniffs, or how big a sniff to take. When subjects were limited to a single sniff, they took one that resembled the first in a natural sniffing episode. Whether the sniff was the first-and-only or the first-of-many, it did not appear to vary with odor strength. After many experiments he could state his findings in a nutshell: “a single natural sniff provides as much information about the presence and intensity of an odour as do seven or more sniffs.” A natural first sniff can’t be beat. (For the technically minded, the optimum sniff has an inhalation rate of 30 liters per minute, a volume of 200 cubic centimeters, and a minimum duration of .40 to .45 seconds.)

There are two aspects to sniffing that are reflected in how we use the verb “sniff.” It can refer to a purely mechanical act (the drawing of air “through the nose with short or sharp audible inhalations”) or to an olfactory experience (“to smell with a sniff or sniffs”). The dictionary’s dichotomy between physical and sensory sniffing is programmed into the central nervous system at a profound level. The brain is not a passive recipient of smells drawn up the nose; it actively manages the acquisition of odor by the nose, and it does so on a time scale of milliseconds.

UC Berkeley smell researcher Noam Sobel was puzzled to find smell-related activity in the cerebellum, a brain area principally involved in tactile discrimination and the control of motor movements. When he and his lab team followed up, they discovered that two parts of the cerebellum were involved in sniffing. One was a smell-activated area; it lit up when a person smelled an odor. The stronger the odor, the greater the activation. Normally this area is activated in the course of sniffing scented air. Sobel found it was also activated by passive smelling, where odors were puffed into the subject’s nose through a tube while they held their breath. The second area of the cerebellum is sniff-activated; it lights up during the physical act of sniffing, but not during passive smelling. The sensation of air flowing through the nose explains the activation in the tactile part of the brain. When topical anesthetic was applied to a subject’s nasal passages to numb the nose, brain activity plunged. Together, two brain areas adjust sniff size to odor strength. This feedback happens very quickly: less than two-tenths of a second into the sniff. (By measuring with far greater precision than was available to Donald Laing, Sobel’s group found that the first sniff of a series was not fixed—only its first 160 milliseconds were.) As a strong odor is detected, the cerebellum signals the respiratory muscles to throttle back on the sniff. What appeared at first to be anomalous brain activity led Sobel and his team to a new understanding of how the brain shapes our perception of smell. The cerebellum is doing what it excels at: monitoring sensory input (in this case odor strength), in order to control a motor action (inhalation).