A controlled experiment has shown that horses respond to the smell of human fear with immediate changes in behaviour and physiology. Researchers collected sweat from people watching a horror scene and placed the samples into cotton pads fitted inside a lycra muzzle that sat directly in front of each horse’s nostrils. Other horses received pads carrying sweat from people watching joyful clips. A third group received unused cotton pads. Forty three mares were tested, each exposed to one scent before entering a structured sequence of behavioural assessments.
In the grooming test, a handler used a standardised method to scratch the wither and neck. Horses exposed to fear scent touched the handler fewer times. Touching the handler is treated as a simple measure of willingness to initiate contact. Lower contact rates aligned with increased caution. Horses exposed to joyful scent or unused pads showed higher interaction levels.
The free human approach test followed. The handler stood motionless for three minutes while the horse was free to initiate contact. Horses exposed to fear scent approached less often and with less confidence. Horses in the joy and control groups made more voluntary contacts. The pattern matched the results from the grooming test, showing the same decline in social engagement when fear scent was present.
Reactivity was then tested with a sudden visual stimulus. A bucket of food was placed on the ground. When a horse began eating, an assistant opened an umbrella near the bucket. Horses exposed to fear scent produced stronger startle responses. The scoring scale included head lifts, stepping back, quarter turns, and rapid half turns. Horses in the fear group consistently hit higher values on this scale. Their maximum heart rate during the startle response was also higher than the other groups. This showed a physiological match to the behavioural intensity seen in the umbrella test.
A novel object test added another dimension. A one-meter object composed of linoleum and plastic shapes was placed in the enclosure. Horses were allowed three minutes to explore it. Horses exposed to fear scent stared at the object for longer periods and made fewer contacts with it. Horses exposed to joyful scent approached more readily and made more direct contact. Horses in the control group fell between the two. Increased vigilance and reduced exploration are well established indicators of elevated arousal, and the pattern repeated consistently in this group.
All behavioural variables were combined using principal component analysis. The main axis represented a continuum between willingness to interact with the handler on the positive end and intensity of fear reactions on the negative end. Horses exposed to fear scent clustered toward the negative side of this axis. Horses in the joy and control groups clustered toward the positive side. This unified model showed that the entire behavioural profile shifted when horses inhaled fear odour.
Heart rate was recorded during the free approach, suddenness test, and novel object test. Artefacts were removed with a strict filter. Maximum heart rate during the suddenness test stood out as the only consistent physiological difference. Horses exposed to fear scent reached higher peaks than either comparison group. Mean heart rate values did not differ, but peak responses matched the intensity seen in the startle behaviours.
Cortisol was measured with saliva samples taken before and after each block of tests. Cortisol changes varied widely among individuals and did not separate the groups. The behavioural and cardiac data provided a clearer signal, showing faster and more direct sensitivity to the scent conditions.
The human sweat samples used in the study were tightly controlled. Volunteers followed dietary restrictions, used unscented hygiene products, and wore cotton shirts and pads washed without fragrance. Sweat was collected while each participant watched either a fear clip or joyful clips. Pads were frozen immediately. Fourteen donors were selected from the larger participant group based on the strongest emotional responses reported on standardised scales. Their pads were used across the horse tests to maintain consistency.
The horses were kept in stable social groups and were accustomed to human handling. Each horse completed a morning set of tests and an afternoon set. The order of tests was fixed from lowest reactivity to highest. The experimenter handling the horse did not know which scent each horse had been assigned. Assistants preparing pads and recording behaviour followed strict routines to limit contamination, including gloves, masks, unscented washing routines, and cleaning procedures between horses.
Across all tests, the same pattern emerged. Horses exposed to fear scent reduced voluntary contact with the handler. They became more reactive to sudden movement. They showed greater vigilance toward new objects. They reached higher peak heart rates when startled. Joy scent did not create these shifts, and unused pads produced neutral responses. The factor that consistently altered behaviour and physiology was exposure to human fear odour.
The shift did not require any visual or auditory cue. The horses reacted to scent alone. The cotton pads were hidden inside the muzzle, out of sight. The environment, handler, and procedures remained identical for all horses, which means the odour itself was the driver.
Human fear sweat contains chemical compounds produced during heightened arousal. Previous research has identified possible contributors, including metabolites linked to the adrenaline pathway. This experiment did not isolate any single molecule, but the behavioural pattern shows that whole scent samples created under fear conditions carry enough information to trigger a measurable change in another species.
Horse training, equine therapy programs, veterinary procedures, and riding environments depend on close human proximity. A horse standing beside a nervous handler can detect chemical cues from human skin before any physical signs of tension become obvious. The study’s behavioural pattern shows that these cues can increase reactivity, reduce cooperation, and produce faster, more intense responses to sudden movement.
Horses use scent in their own social interactions. The ability to detect human emotional odour fits within their existing sensory framework. Chemical cues operate without sound, without movement, and without line of sight. In controlled conditions, they were strong enough to shift both behaviour and physiology. The findings show that horses register human emotional states through scent and respond immediately.
Source:
Jardat P. et al. 2026. Human emotional odours influence horses’ behaviour and physiology. PLOS ONE.
Link: https://doi.org/10.1371/journal.pone.0337948
