Fear Has A Smell, and Horses Know It

THE horse stands quietly as the experimenter approaches, a cotton pad stapled inside a lycra muzzle fitted over her nostrils.

She’s breathing normally through the fabric, but something is different this time. When the experimenter extends a hand during the grooming test, the mare keeps her distance, touching the human significantly less than she did yesterday. Later, when an umbrella suddenly snaps open beside her, she startles violently; far more than seems warranted for a horse accustomed to the test.

The mare doesn’t know it, but those cotton pads carry molecules of human fear. Somewhere in the chemistry of sweat, terror has left its signature.

We’ve long known that animals communicate through smell, from ants laying pheromone trails to wolves marking territory with urine. But most research has focused on signals within a species, particularly around reproduction and danger. The question of whether emotional information can cross the species barrier, transmitted through odor alone, has remained largely unexplored. Until now.

Researchers at INRAE and the French Horse and Riding Institute have demonstrated that horses don’t just detect human emotional odors; they respond with matching emotional states. When exposed to sweat collected from people watching horror films, horses exhibited heightened fear responses and avoided human contact. The finding suggests emotional contagion through chemical cues, a phenomenon previously documented only in dogs among interspecies pairs.

The experimental design was straightforward, though hardly simple to execute. Léa Lansade and her colleagues recruited 30 volunteers who agreed to some unusual viewing assignments. On two different days, each participant watched either a 20-minute excerpt from the horror film Sinister or a montage of comedy sketches and musicals designed to elicit joy. During the screenings, cotton pads tucked under their armpits collected whatever their bodies were secreting.

The researchers weren’t just hoping for sweat. Human apocrine glands, concentrated in armpits, produce compounds like adrenaline and androstadienone that seem to carry emotional information, at least among humans. More than 40 studies have shown that these “chemosignals” can influence the emotional state of other people. Fear-related sweat, for instance, can trigger fearful facial expressions in those who smell it. Whether horses possess receptors tuned to these same molecules remains unknown.

After the volunteers completed their sessions, the research team selected samples from the 14 participants who reported feeling the most fear during the horror film and the most joy during the comedy clips. These pads, along with unused control pads, were frozen at −70°C until testing day.

Then came the horses. Forty-three Welsh mares, all accustomed to human handling, were divided into three groups. Each would smell either fear-related odors, joy-related odors, or control odors, but not until right before a series of behavioral tests designed to measure both fear responses and willingness to interact with humans.

The protocol was carefully controlled. The experimenter who conducted the tests didn’t know which group each horse belonged to, and an assistant recorded behaviors from outside the test box. Before each session, fresh cotton pads were thawed and stapled into a disposable lycra muzzle that fit over the horse’s muzzle, positioning the pads directly in front of the nostrils. The horses could breathe normally. They just couldn’t escape the scent.

The results were striking. During the free human approach test, horses exposed to fear odors touched the experimenter 40 percent less than those smelling joy odors. In the suddenness test—where an umbrella opened abruptly while the horse ate from a bucket—fear-scented horses startled more intensely, and their heart rates spiked higher than both control and joy groups. When presented with an unfamiliar object (a colourful assemblage of linoleum and plastic pieces), they gazed at it more frequently, a recognized indicator of anxiety in horses.

Principal component analysis confirmed the overall pattern: horses smelling fear-related odors were significantly more fearful and less willing to approach humans compared to the other groups. Salivary cortisol measurements showed high variability, probably due to the hormone’s circadian rhythms, but heart rate data from the suddenness test left little doubt. The fear odors triggered physiological arousal.

More than a simple reaction to an unfamiliar smell, the horses appeared to adopt the emotional state itself. This matches the criteria for emotional contagion, where an emitter’s emotional state transmits to a receiver. The correspondence in both valence (negative emotion) and arousal (heightened state) between the human sweat donors and the horses suggests the animals weren’t just reacting—they were catching human fear.

Why might this ability have evolved? Fear signals danger. For species that regularly interact, perceiving a companion’s fear could be adaptive even across species boundaries. If your human handler is afraid, perhaps you should be too. Olfactory signals are particularly effective because they work in darkness, persist after the emitter has left, and penetrate obstacles. In the wild, recognizing fear in another species might mean the difference between staying put and fleeing a predator.

Interestingly, horses exposed to joy-related odors didn’t behave significantly differently from those smelling unused cotton pads. This might mean joy signals are more subtle, or that positive chemosignals are simply less salient than negative ones. Fear motivates immediate action; joy does not. Another possibility: “joy” is too complex an emotion to transmit clearly across species, while fear, rooted in the most ancient survival circuits, translates more universally.

The researchers acknowledge several limitations. The study used only female Welsh ponies, all between 5 and 12 years old, so responses in other breeds, ages, or sexes might differ. An audience horse was present during testing to prevent social isolation stress, which could have influenced results despite precautions. And cortisol measurements were inconclusive, probably because the tests were too brief to capture hormonal changes that unfold over hours.

Still, the findings raise intriguing questions about domestication. Most animals that respond to human emotional signals are domestic species: horses, dogs, cats, cattle, goats. Is this capacity something domestication selected for, or did humans preferentially domesticate species whose ancestors already possessed it? Dogs were previously the only species shown to respond to human fear odors, but horses have now joined that exclusive club.

“There is a chemical communication of emotions, and it crosses the species barrier,” says Lansade. The study, she notes, shows that horses exposed to human fear odors exhibit fear responses through emotional contagion.

The practical implications are immediate. In stables, training arenas, and therapeutic riding programmes, human emotional states might directly affect horse behavior and welfare. A nervous rider doesn’t just transmit tension through the reins; they might be broadcasting chemical distress signals. Handlers experiencing stress or fear could inadvertently increase horse reactivity, potentially raising accident risk or undermining training effectiveness.

This doesn’t mean handlers need to maintain perfect calm at all times. But it does suggest that emotional regulation, often overlooked in equestrian training, deserves more attention. Perhaps handlers could be selected or trained based partly on their emotional consistency, or strategies developed to minimize stress transmission during high-stakes interactions.

We’re only beginning to understand how deeply interspecies communication runs. The nose, it turns out, may be as important as the eyes and ears in reading the emotional landscape around us. And what horses smell when we approach them might matter far more than we ever imagined.