Skip to content
NatureHQ

Senses and abilitiesability

Smell, and how far it really works

A moth can detect a single molecule and still cannot fly straight to the source. The nose was never the limiting factor.

Detecting a substance and finding its source are two different problems, and almost every dramatic claim about animal smell confuses them. A moth detects single molecules and still cannot fly straight to a female — because a plume is a broken, shifting thing, and following one means solving a navigation problem in it.

Olfaction is where the gap between a laboratory threshold and a real capability is widest, and where the folklore is thickest as a result. The thresholds are genuinely extraordinary: a single pheromone molecule can produce a recordable response in a single receptor on a moth’s antenna, and dogs detect some compounds at concentrations that are difficult to produce reliably. None of that tells you how far away an animal can locate something, because the limiting factor is not the nose. Odour in air or water does not spread as a smooth cone with the strongest concentration in the middle; turbulence tears it into filaments, and an animal inside a plume meets brief bursts separated by nothing, with the gaps carrying as much information as the hits. So tracking is a behavioural strategy rather than a matter of following a gradient — surge when you get a hit, cast across the flow when you lose it, and use the timing rather than the strength of what arrives. That is what the moth is doing, and what the shark is doing, and the shark result makes the point sharply: given the same odour at both nostrils with a tiny delay between them, the animal turns towards the side that got it first, and it does that even when the other side got a stronger dose. Timing beats concentration. And once you know that, "smelling blood from miles away" stops being an exaggeration of a real capacity and becomes a claim about the wrong quantity altogether.

Developed coverage · 72% complete · reviewed 2026-09-03

What this page covers

Chemical sensing is close to universal. The material here is drawn mainly from work on sharks, moths, dogs and moles — the systems where tracking behaviour has been measured rather than inferred from receptor counts.

Often confused with: Sensitivity, which is how little of a substance can be detected, and which does not determine how far away an animal can find something

Quick facts

The real problem
Not detection — finding the source of a plume that turbulence has torn apart
What a shark steers by
Which nostril got the odour first, not which got more of it
Blood from miles away
A claim about the wrong quantity
Smelling underwater
Some mammals do it by exhaling a bubble onto a scent and re-inhaling it

Where this appears

Every organism below has been linked to this page because the evidence links them. Each one carries its own evidence, and its own limits.

A plume is not a cone

The picture everyone has of how smell spreads is wrong in a way that changes the whole problem.

Diagram

What an animal inside a plume is actually dealing with

Schematic. Filament structure is drawn far more orderly than a real plume.

Why following a smell is not walking up a gradientsourceA real plume is torn into filaments. At any instant, most of it is empty.What animals actually doDetect something: head upwind. Lose it: cast across the wind until you find itagain. A shark adds a trick: turn towards whichever nostril smelled it first.Schematic. Filament positions are illustrative, not measured.
The same explanation in words

Two panels. The upper, marked as the intuitive picture and struck through, shows a smooth widening cone spreading from a source, strongest along the centre line — a gradient an animal could climb. The lower, marked as what is really there, shows the same source producing a ragged set of separated filaments that meander and break apart downstream, with wide empty gaps between them. Beneath, the tracking strategy is set out as two moves: on contact with a filament, surge upwind or upcurrent; on losing it, cast across the flow until another is found. A closing note says the information is in the timing of the encounters rather than in their strength.

A male silk moth’s antenna responds to a single molecule of pheromone. That is the physical limit of sensitivity, and it is not what lets him cross a field to find a female — the flight strategy does that.

Established

Specialists would state this without hedging. Multiple independent lines of evidence agree.

Individual sensilla on the silk moth antenna respond to single pheromone molecules, with behavioural thresholds of a small number of captures within a short interval. Long-range localisation depends on a search strategy — upwind surge on contact, crosswind casting on loss — because a plume at distance consists of intermittent filaments carrying no usable concentration gradient.

Who this applies to
One exceptionally well-studied pheromone system; the surge-and-cast strategy is documented across many flying insects.
Studied in
Bombyx mori
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

The receptor sensitivity is measured directly from single cells, and the plume physics that makes a gradient unusable is well characterised.

How far it can be extended

Surge-and-cast plume tracking is observed widely in moths and other flying insects following odour.

Caveats

  • Threshold measurements under controlled airflow do not translate into a field range.
  • One species and one pheromone system, both unusually well characterised.
  • Reported tracking distances vary enormously with wind and vegetation.

Still unanswered

  • The maximum distance over which wild males actually locate females, as opposed to the distances quoted from release experiments.

Last reviewed 2026-09-02

The evidence (2 studies)

How we know

Recording from one hair on a moth’s antenna

How little of a scent does a moth need in order to detect it?

A silk moth’s antenna is covered in thousands of fine hairs, each containing sensory cells. An electrode was placed against a single hair so that the electrical response of the cells inside it could be recorded, and controlled quantities of the female pheromone were delivered in a measured airstream. The amount was reduced until the responses stopped, and separately the behavioural threshold — the amount needed before a male begins his search flight — was measured.

What happened

A single sensory cell responds to the capture of a single molecule of pheromone, and a male begins searching after a small number of such captures across the antenna within a short interval.

What it shows

Detection at the physical limit — you cannot be more sensitive to a chemical than one molecule. The antenna is essentially a comb built to intercept as much passing air as possible with as many detectors as possible.

What it does not show

It does not show that a moth can smell a female from kilometres away, which is the claim it is usually cited for. Single-molecule sensitivity does not deliver a direction: at that distance the pheromone arrives as isolated filaments with long gaps between them, and the moth covers the ground by flying upwind while it detects anything and casting crosswind when it loses the trail. The distance comes from the search strategy, not from the threshold.

The controls — what makes this evidence rather than a story
  • Recordings from single identified cells rather than the whole antenna, so the response is attributable.
  • Delivered doses measured rather than estimated, with the airflow controlled.
  • Related compounds tested alongside, establishing that the cells are specific rather than generally excitable.

From R. H. Wright Lectures on Insect Olfaction

The moth work is the origin of both halves of this. Recording from one sensory hair established a sensitivity that is still startling — and the same body of work established that an antenna is built to resolve rapid changes rather than to integrate a concentration, which only makes sense if what matters is the pattern of arrivals. A nose optimised for a gradient would do the opposite.

Words used here
Odour plume
The region downwind or downcurrent of a scent source. Turbulence breaks it into filaments, so it is patchy rather than a smooth gradient.
Casting
Sweeping across the flow after losing contact with a plume, to cut back across it. The counterpart to surging upwind on contact.

How far can an animal really smell something?

The most common question about this sense, and the one where the popular answer is a category error.

How we know

Delivering a smell to one shark nostril a fraction of a second before the other

A shark following a scent has to decide which way to turn. Is it comparing how strong the smell is on each side?

Sharks were fitted with headgear that delivered odour to each nostril through a separate tube, so that the two sides could be controlled independently. The experiment then varied two things separately: which nostril received the odour *first*, and by how much, and how *strong* the odour was on each side. A shark comparing concentration should turn towards the stronger side. A shark comparing timing should turn towards the earlier one.

What happened

The sharks turned towards whichever nostril smelled it first, down to delays of a fraction of a second, and did not turn towards the stronger side.

What it shows

That smell-tracking works like sound localisation rather than like following a gradient. This matters because a real plume is not a gradient at all — it is a turbulent tangle of filaments in which the concentration at any instant says almost nothing about direction, while the order of arrival still does. It explains how an animal holds a heading in water where the obvious cue is noise.

What it does not show

It shows how a shark steers once it is in a plume, not how it finds one, and says nothing about the distance over which any of this works — which is the part of the popular claim that most needs testing. The animals were restrained and instrumented, which constrains natural swimming, and concentration may still matter over longer timescales.

The controls — what makes this evidence rather than a story
  • Timing and concentration varied independently, which is the only way to tell the two hypotheses apart.
  • Both nostrils instrumented identically, so the apparatus cannot produce a bias.
  • Delays reduced step by step to find the threshold rather than testing one interval.

From The function of bilateral odor arrival time differences in olfactory orientation of sharks

A shark following a smell turns towards whichever nostril detected it first, not towards the side where it is strongest — because in a real plume, strength tells you almost nothing.

Well supported

Good evidence backs this, though some details remain open.

Smooth dogfish turn towards the naris receiving an odour pulse earlier, at inter-naris delays down to fractions of a second, and do not orient according to concentration differences between the nares. This is consistent with plume structure, in which instantaneous concentration is spatially uninformative under turbulent mixing.

Who this applies to
Demonstrated in one shark; bilateral timing comparison is proposed more widely in aquatic olfaction.
Studied in
Mustelus canis
Why we rate it this way, and what the caveats are
Well supportedHigh confidence

Timing and concentration were varied independently, which is the only design that can separate the two hypotheses, and the result is unambiguous.

Caveats

  • Shows how a shark steers once inside a plume, not how it finds one.
  • The animals were restrained and instrumented, which constrains natural swimming.
  • Concentration may still matter on longer timescales than this design tested.

Still unanswered

  • Whether bilateral timing comparison is general across aquatic olfaction or specific to animals with widely separated nares.

Last reviewed 2026-09-02

The evidence (2 studies)

Sharks smell well, and use it to navigate — not to detect a drop of blood from miles off

Well supported

Good evidence backs this, though some details remain open.

Shark olfactory thresholds are comparable to those of other fishes, in the parts-per-billion range for some compounds. Odour reaches a shark only as fast as the current carries it, and the demonstrated navigational role operates over kilometres and hours. Sharks with olfaction blocked and displaced offshore still returned to shore, but by markedly less direct routes.

Who this applies to
demonstrated in coastal sharks; olfactory physiology measured across several species
Studied in
Triakis semifasciata, Carcharhinus plumbeus

You may have heard

“A shark can smell a single drop of blood from miles away”

Two errors in one sentence. Shark olfactory thresholds are good but not extraordinary among fishes — roughly one part per billion for some compounds, not one drop in an ocean. And detection depends on the molecules arriving: a shark upstream of blood smells nothing at all, however close, while a plume takes hours to travel a mile with the current.

Why we rate it this way, and what the caveats are
Well supportedHigh confidence

The positive finding — that smell contributes to navigation over kilometres — comes from a displacement experiment with a sensory-blocking control. The negative half rests on measured thresholds and on the physics of odour transport in water, neither of which is in dispute.

How far it can be extended

Olfactory receptor physiology is broadly similar across sharks, and the transport physics that limits odour detection applies regardless of species.

Caveats

  • Sharks genuinely do have excellent olfaction; the correction is about range and about what the sense is for.
  • A shark can follow an odour plume for a long distance — but only downstream of it, and only as fast as the water moves.
  • Sensitivity varies by compound and species; blood is not a special case.

Still unanswered

  • Which chemical gradients are being used for navigation?
  • How do sharks combine olfactory, magnetic and visual cues on long migrations?

Last reviewed 2026-08-09

The evidence (2 studies)

Put the two results together and the popular claim dissolves for a specific reason rather than a general one. The distance a shark can find something from is set by whether a plume reaches it in a followable state and how long it takes to work upstream along it — a question about currents and about search behaviour. Detecting a dilute trace at a great distance would not help if the trace arrives with no directional information, and that, not the sensitivity of the nose, is the binding constraint.

Smelling underwater with a nose built for air

How we know

Filming a mole blowing bubbles at things underwater

Mammals are supposed to be unable to smell underwater, because smelling requires air. So how does a semi-aquatic mole follow a scent trail in a stream?

Star-nosed moles and water shrews were filmed foraging underwater with a high-speed camera — at ordinary frame rates the relevant behaviour is invisible. The animals were given scent trails to follow, and objects both scented and unscented to investigate. The bubble behaviour was then prevented, and trail-following performance measured again.

What happened

The animals exhaled small air bubbles onto objects and the substrate and re-inhaled them, up to about ten times a second, and followed scent trails successfully. Preventing the bubbles impaired their performance.

What it shows

That the premise was wrong in an unexpected way. Smelling does require air — so the animal takes air with it, presses it against the thing it wants to smell, and breathes it back in. It is also a demonstration of how much sensory biology depends on the frame rate of the camera: at normal speed there is nothing to see.

What it does not show

Two species with unusual foraging ecology, neither of them representative of mammals; this is not a general mammalian ability. It establishes that odour is transferred and used, not how much information is recovered, and the work was done in tanks rather than in a stream.

The controls — what makes this evidence rather than a story
  • High-speed film, without which the behaviour cannot be seen at all.
  • Scented and unscented targets, so a response can be attributed to odour.
  • Trials with the bubble behaviour blocked, testing whether it is doing the work or merely accompanying it.

From Underwater "sniffing" by semi-aquatic mammals

Smelling needs air, so a star-nosed mole takes air with it: it blows a bubble onto whatever it wants to smell and breathes it back in, about ten times a second.

Well supported

Good evidence backs this, though some details remain open.

Star-nosed moles and water shrews exhale and re-inhale air bubbles against submerged substrates and objects at rates up to approximately 10 Hz, and follow submerged scent trails. Preventing the behaviour impairs trail-following performance.

Who this applies to
Two semi-aquatic insectivores; not a general mammalian capacity.Do not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Condylura cristata, Sorex palustris
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

The behaviour is unambiguous on high-speed film and the blocking control ties it to performance. It rests on one study in two species, which is why confidence is not higher.

How far it can be extended

Both species have unusual semi-aquatic foraging ecology, and the behaviour has not been found in mammals generally.

Caveats

  • Two species with unusual ecology; this is not something mammals generally do.
  • Laboratory tanks rather than natural streams.
  • Establishes odour transfer and use, not how much information is recovered.

Still unanswered

  • Whether other semi-aquatic mammals do the same thing and nobody has filmed them fast enough.

Last reviewed 2026-09-02

The evidence (1 study)

This is a good example of a finding that existed as a curiosity in high-speed footage before anyone realised what they were looking at. A mammal’s olfactory epithelium needs air; a submerged mammal has none to spare. Exhaling a bubble onto an object, letting volatile compounds move into it, and inhaling it back is a way of sampling water with an air-based nose — and it is fast enough that the animal does it several times a second. That two unrelated semi-aquatic mammals both do it suggests it is a general solution rather than a quirk.

Where else chemistry does the work

The most famous case of smelling a way home

  • Salmon

    Learned as a smell during a juvenile window — and only the last few kilometres

  • Homing pigeons

    Where an olfactory map is claimed, and argued about

What is still unsettled

  • How much of a dog’s tracking ability is the nose and how much is the search strategy?

    Why it matters: Dogs are the most-used working olfactory animal in the world, and the split between sensitivity and behaviour has direct consequences for how they are trained and how their results should be weighed.

    What would settle it: Tracking experiments where plume structure is measured directly rather than assumed, which is now technically possible and rarely done.

  • Do animals identify a source, or only follow the plume until they arrive at it?

    Why it matters: Discrimination between odours is well demonstrated. Whether a tracking animal knows what it is following before it gets there is a different question and much less studied.

    What would settle it: Choice experiments in which two plumes from different sources overlap, forcing an identification mid-track.

  • How many receptor genes an animal has, and what that predicts

    Why it matters: Receptor gene counts are used constantly as a proxy for olfactory ability, in the same way photoreceptor counts are misused for vision, and the relationship to behaviour is weak.

    What would settle it: Behavioural threshold and tracking measurements across species with well-characterised receptor repertoires, which exist for very few animals.

The research behind this page

11 studies, newest first. Each one has a page explaining what it found and what it could not show.

2016PLoS ONE

Olfaction contributes to pelagic navigation in a coastal shark

Sharks with an intact sense of smell headed shoreward in a comparatively direct line; sharks with olfaction blocked took much more tortuous paths and ended significantly further from shore.

2014PLoS ONE

Multisensory integration and behavioral plasticity in sharks from different ecological niches

Each sense dominates a different phase, and the sequence differs by species.

2013Journal of Experimental Biology

Forty years of olfactory navigation in birds

Anosmic pigeons released at unfamiliar sites are impaired in choosing a homeward bearing, while their compass orientation and their homing from familiar areas remain intact; the effect has been reproduced across laboratories and methods of olfactory deprivation.

2010Current Biology

The function of bilateral odor arrival time differences in olfactory orientation of sharks

Sharks turned towards the side that received the odour first, for delays of roughly a tenth of a second down to a fraction of that, and did not respond to differences in concentration between the two sides.

2008Journal of Experimental Biology

Sensory ecology on the high seas: the odor world of the procellariiform seabirds

Several procellariiform species are attracted to dimethyl sulphide, a compound released where plankton are grazed, and the compound is concentrated over predictable seafloor features rather than being uniformly distributed.

2006Nature

Underwater "sniffing" by semi-aquatic mammals

The animals exhaled air bubbles onto objects and re-inhaled them at rates of up to about ten times a second, and followed scent trails underwater successfully; performance dropped when bubble behaviour was prevented.

2004Animal Behaviour

Avian olfactory navigation: its empirical foundation and conceptual state

Measured atmospheric trace-gas ratios do vary spatially in a sufficiently orderly way to be usable in principle, and the behavioural record is consistent with birds learning such gradients at the home loft.

2004Proceedings of the National Academy of Sciences

Familiar route loyalty implies visual pilotage in the homing pigeon

Individual birds converged on idiosyncratic, highly repeatable routes that were not the straight line home, and clock-shifted birds followed their established routes rather than departing at the deflected bearing.

1987Simon Fraser University

R. H. Wright Lectures on Insect Olfaction

A single receptor cell responds to the capture of a single pheromone molecule, and behavioural responses follow from a small number of such captures across the antenna within a short interval.

1976Science

Imprinting to Chemical Cues: The Basis for Home Stream Selection in Salmon

Salmon exposed as smolts returned overwhelmingly to the stream scented with that chemical, while unexposed controls did not.

1951The American Naturalist

Discrimination of Stream Odors by Fishes and Its Relation to Parent Stream Behavior

Fish could be trained to tell the two waters apart, and lost the ability when their sense of smell was blocked.

This page is a stop on a longer route

A guided journey reads several subjects in a deliberate order, with an argument for why one follows another. You can join in the middle.

Where to go from here

Each of these follows from something on this page — a relationship in the evidence, a claim people ask about, or the next mechanism along.

How complete this page is, and what it is still missing

NatureHQ publishes its own gaps. This page is at 72% completeness against what we would call a finished subject, and was last reviewed on 2026-09-03. It carries 7 claims and answers 13 mapped search questions.

  • no research from the last few years is attached — check for newer work
  • no popular claim about this subject has been checked yet
  • The vomeronasal organ and pheromone detection in vertebrates are named rather than treated, and both carry a great deal of popular misinformation.
  • Taste is not covered here at all, despite being the other chemical sense and being continuous with smell in aquatic animals.
  • Nothing here covers olfactory memory, which is where much of the work on dogs and on rodents actually sits.