Skip to content
NatureHQ

higher taxon

Ants

Formicidae Latreille, 1809

Ants are social insects in the family Formicidae, with over 14,000 species. Much of what looks like ant intelligence belongs to the colony rather than to any individual ant.

Ants are the best argument that intelligence and cleverness are different things. An ant colony finds the shortest route to food within minutes, and no ant compares the routes — the shorter path simply accumulates pheromone faster, so the colony behaves as though it decided. Leafcutters run fungus farms with bacterial crop protection, a system tens of millions of years older than agriculture. Individual ants do have real abilities: a desert ant measures distance by counting its own strides, which was established by gluing stilts to their legs. And one widely shared claim — that ants pass the mirror test — rests on a study that does not support it, which is why it has its own entry here.

Developed record · 97% complete · reviewed 2026-08-09

What this page covers

Over 14,000 described species with wildly different lifestyles. A finding about Saharan desert ants tells you very little about the ants in a British kitchen.

Often confused with: Isoptera (termites — often called "white ants", but they are closer to cockroaches)

Quick facts

Species
Over 14,000 described
Distance sense
Saharan desert ants integrate stride counts
Farming
Leafcutters cultivate fungus and carry antibiotic-producing bacteria
Neurons
Around 250,000 in a typical worker

What a colony actually is

Nobody is in charge, least of all the queen.

The queen lays eggs. She does not give orders, and nothing else does either

Established

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

An ant colony is a family group in which the queen’s role is reproductive. Task allocation arises from workers responding to local chemical and tactile cues, thresholds that differ between individuals, and the colony’s current needs — not from direction by the queen or by any central process.

Who this applies to
ants generally, with wide variation in queen number and colony structure
Studied in
Formicidae

You may have heard

“The queen ant commands the colony”

The word arrived with a monarchy attached. She is a reproductive organ with legs, and the colony organises itself from thousands of workers following local rules — which is a stranger and better answer than a chain of command.

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

A foundational and undisputed result in social insect biology, supported by monographic synthesis and by direct experimental work on trail-based foraging.

How far it can be extended

Decentralised task allocation is documented independently across many ant subfamilies; the reproductive role of queens is a defining feature of the family.

Caveats

  • Colony structure varies enormously: some species have one queen, some many, some none at any given time.
  • Queens do influence the colony chemically — suppressing worker reproduction — which is regulation, not instruction.
  • Decentralised does not mean unstructured; workers differ systematically in response thresholds.

Still unanswered

  • What determines an individual worker’s response threshold?
  • How do colonies of very different sizes keep the same task ratios?

Last reviewed 2026-08-10

The evidence (3 studies)

A colony is a family. The queen is the mother of nearly everyone in it, and her job is laying eggs — she issues no instructions and would have no way to. Workers are her sterile daughters. Males exist briefly, mate once and die, and are produced only when the colony is ready to send out reproductives.

The word "queen" arrived carrying a monarchy that is not there, in the same way "alpha" imported a hierarchy into wolves. What actually allocates work is far stranger: individual ants differ in how readily they respond to a given cue, so a rising smell of hungry larvae recruits nurses and a fallen trail recruits foragers, and the ratios come out roughly right without anything computing them.

Who is who
CasteWhat it doesHow long it lives
QueenLays eggs; founded the colony after a mating flightYears to decades in many species
WorkerEverything else — brood care, foraging, defence, wasteWeeks to months
MaleMates on a nuptial flightDays to weeks

Caste is not genetic in most ants: a female larva becomes a queen or a worker according to how she is fed. The colony decides what to build by how it feeds its young.

Words used here
Caste
A role in a colony with a distinct body form. In most ants it is determined by larval diet rather than by genes.
Nuptial flight
The mass emergence of winged reproductives on a single day. Why flying ants all appear at once.

How an ant experiences the world

Mostly smell, some sight, no lungs and no ears as such.

An ant lives in a chemical world. Its antennae are its primary sense organs, detecting both smell and touch, and nearly everything a colony does — recognising a nestmate, following a trail, raising an alarm, tending brood — runs on chemistry rather than on sight or sound.

The questions people ask about ant bodies
FeatureWhat is actually true
EyesMost ants have compound eyes and see movement and shape poorly. Some subterranean species are effectively blind; jumping and trap-jaw ants see very well.
EarsNone. Many ants detect vibration through their legs, and some stridulate — squeak by rubbing body segments — to signal distress.
LungsNone. Air enters through spiracles along the body and reaches tissues through tubes called tracheae.
WingsOnly the reproductives, and only briefly. Queens and males grow wings for a mating flight, after which the queen tears her own off.
Bite or stingBoth, depending on species. Ants bite with mandibles; many also sting, and some spray formic acid instead.
StrengthAn ant can lift many times its own weight because muscle force scales with cross-section while weight scales with volume. Being small is the trick, not the muscle.
DrowningAnts can survive some time submerged and some species raft together on water. They are not amphibious.

The nestmate-recognition system is the one worth understanding, because it explains the colony. Ants carry a blend of cuticular hydrocarbons — waxy compounds on the outside of the body — that is characteristic of their colony. An ant smelling wrong is attacked. A colony is held together by a shared smell rather than by anyone knowing anyone.

Words used here
Cuticular hydrocarbon
A waxy compound on an insect’s surface. The blend acts as a colony badge, and getting it wrong is fatal.
Trachea
A tube carrying air directly to tissues. It works well at insect scale and is one reason insects cannot be large.

Ants in the house, and ants in the garden

Ants come indoors for the same reason they go anywhere: a scout found something and laid a trail home. That is also the weakness. The line of ants on a worktop is following a chemical path, and washing the surface with detergent removes the path — which does more than killing the ants walking on it, because the trail is what keeps bringing more.

Practical

Working with the mechanism rather than against it

Remove the food source and wash the trail rather than treating each ant. Sugar, grease and pet food are the usual attractants, and a sealed container ends the problem faster than any spray. Ants entering a house are rarely damaging — carpenter ants, which excavate wood, are the exception and are worth identifying properly. NatureHQ does not recommend specific pesticides; where control is genuinely needed, product approval and safe use vary by country and a local professional is the right source.

Where this applies: General household guidance; pesticide rules and products differ by country.

In a garden ants are mostly useful: they aerate soil, remove pest insects and disperse seeds — many woodland plants have seeds with fatty attachments that exist specifically to get ants to carry them. The exception is aphid farming. Ants defend aphids for their honeydew, so a plant covered in ants usually has an aphid problem underneath, and the ants are a symptom.

The peony story is folklore. Peony buds produce nectar and attract ants; the flowers open perfectly well without them.

Words used here
Myrmecochory
Seed dispersal by ants. The seed carries a fatty body the ants want; the seed gets planted as a side effect.

How ants find food, and how they lose it

A foraging ant that finds something good lays a chemical trail on the way home, and other ants that stumble onto it follow and reinforce it. Nothing communicates where the food is; the trail is the message and the message is simply "this way was worth walking".

Because the chemical evaporates, the whole system is self-correcting. A trail to exhausted food stops being reinforced and fades within minutes to hours, which is why a marching line of ants disappears so abruptly once the spill is cleaned up — and why wiping the surface, which removes the trail, works better than killing the ants on it.

The textbook picture of an ant as a pure trail-follower is too simple, though. Experienced foragers often prefer their own memory of a route over the trail, and the balance shifts with how reliable each has proved. A colony is not trapped by a stale trail because a proportion of its ants are ignoring it.

  • Most ants are omnivores: sugars from honeydew and nectar for the adults, protein for the larvae.
  • Aphid-farming ants defend aphids and harvest their honeydew, a genuine agricultural relationship.
  • Leafcutters eat neither leaves nor the ants that carry them — they feed the leaves to a fungus and eat that.
  • Trails and repellent "no entry" markers coexist; a colony marks unproductive branches as well as good ones.
Words used here
Trail pheromone
A chemical laid on the ground that other ants follow and reinforce. It evaporates, which is what makes the system self-correcting.
Honeydew
The sugary waste aphids excrete, and a staple food for many ants.

The colony is the clever one

Shortest-path finding, with no ant doing any finding.

An ant colony finds the shortest route without any ant comparing the options

Established

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

In a double-bridge apparatus, Linepithema humile traffic converges on the shorter branch within minutes through pheromone reinforcement: ants using the shorter route return sooner and reinforce it more frequently.

Who this applies to
Argentine ants, in a laboratory apparatus
Studied in
Linepithema humile
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

Repeatedly replicated, with a mathematical model that predicts the outcome quantitatively — and a well-known engineering application built on it.

How far it can be extended

Trail-reinforcement dynamics have been demonstrated in several strongly trail-laying ant species and are well described mathematically. Species that forage individually rather than by trail do not behave this way.

Caveats

  • Colonies do not always find the best route: change the geometry, or add a shortcut after a trail is set, and they can stay on a worse one.
  • This is a property of the colony, not of any ant.

Still unanswered

  • How do colonies escape an established but suboptimal trail in the wild?

Last reviewed 2026-08-09

The evidence (1 study)

How we know

The double bridge

How does an ant colony find the shortest route to food when no individual ant can compare the options?

A nest was connected to a food source by a bridge that split into two branches and rejoined — one branch noticeably longer than the other. Ants could take either. Researchers simply recorded which branch the traffic used as the minutes passed.

What happened

Within minutes, almost all traffic was on the shorter branch — even though no ant had walked both and compared them.

What it shows

The colony solved the problem without anyone solving it. Ants on the short branch get back sooner, so they lay pheromone over that route more often, which attracts more ants, which lays more pheromone. The shortest path simply accumulates the strongest signal first.

What it does not show

It does not mean ants are individually clever, or that colonies always find the best route — change the geometry, or introduce the shortcut after a trail is established, and the colony can get stuck on a worse path.

The controls — what makes this evidence rather than a story
  • Branches were otherwise identical in width and surface.
  • Trials were repeated with the long and short branches swapped, ruling out a side preference.
  • Equal-length branches gave a baseline in which no route should win.

From Self-organized shortcuts in the Argentine ant

This experiment gave computer science "ant colony optimisation", a family of algorithms now used for real routing and scheduling problems.

Words used here
Stigmergy
Coordination through traces left in the environment rather than through direct communication. Pheromone trails are the classic example.

Ants on stilts

Desert ants measure how far they have walked by counting their steps

Established

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

Cataglyphis fortis foragers whose legs were lengthened with stilts overshot the nest on the return journey, and those with shortened legs stopped short. Ants that made both journeys on modified legs returned accurately, indicating a stride-based odometer.

Who this applies to
Saharan desert ants
Studied in
Cataglyphis fortis
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

An unusually decisive manipulation with a control condition that eliminates the obvious alternative explanation, and the predicted direction of error in each group.

Caveats

  • One desert species navigating in terrain with almost no landmarks.
  • Stride integration is one component alongside path integration and sky-compass cues.
  • "Counting" here means integrating stride signals, not deliberate enumeration.

Still unanswered

  • How is the stride signal actually accumulated and stored?

Last reviewed 2026-08-09

The evidence (1 study)

How we know

Ants on stilts

How does a desert ant know how far it has walked, in a landscape with almost no landmarks?

Saharan desert ants were trained to walk a set distance from their nest to a feeder. At the feeder — after the outward journey was already complete — some ants had lengths of pig bristle glued to their legs as stilts, and others had the ends of their legs trimmed. Then they were allowed to head home. If an ant measures distance by counting strides, a longer stride should carry it too far and a shorter one should leave it short.

What happened

Stilted ants marched straight past the nest and began searching too far away. Ants with shortened legs stopped short and searched too early. The group that walked both ways on modified legs got it right.

What it shows

Desert ants estimate distance from the number of strides they take. The group that made both journeys on modified legs is what clinches it: they recalibrated, so the mechanism is counting strides, not sensing leg length.

What it does not show

It does not mean ants count in any deliberate sense, and it does not describe how other ants navigate. This is one desert species that lives where there is nothing to look at; ants in cluttered habitats lean far more on landmarks and pheromone trails.

The controls — what makes this evidence rather than a story
  • Unmodified ants walked the same route to establish the correct homing distance.
  • A second group was modified *before* the outward journey, so they walked out and back on the same legs.
  • The terrain was flat and featureless, removing landmark cues.

From The ant odometer: stepping on stilts and stumps

Compare across species

Ants that teach

Some ants teach each other routes, by a formal definition of teaching

Well supported

Good evidence backs this, though some details remain open.

In Temnothorax albipennis, a leading ant adjusts its pace to the follower's contact, incurs a cost in slower travel, and the follower subsequently reaches the destination faster than it would alone — satisfying the standard functional criteria for teaching.

Who this applies to
one small ant species that moves nest deliberately
Studied in
Temnothorax albipennis

You may have heard

“Ants teach each other, just like humans do.”

The definition being met is a behavioural one written by biologists: a cost to the teacher, adjustment to the pupil, and faster learning as a result. It does not imply that the leading ant knows the follower is ignorant, or intends to help.

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

The three criteria were each measured directly: bidirectional feedback in pace, a measured cost to the leader, and improved performance by the follower.

Caveats

  • This meets a functional definition. It says nothing about intention or understanding.
  • One species with unusually deliberate nest-moving behaviour.

Still unanswered

  • Do other tandem-running species meet the same criteria?

Last reviewed 2026-08-09

The evidence (1 study)

  • Supports · primary

    Teaching in tandem-running ants

    Franks and Richardson, 2006 · Nature

    Leaders slowed and accelerated in response to the follower, at a cost to their own journey time.

The definition being satisfied is a formal one: the teacher pays a cost, adjusts to the pupil, and the pupil learns faster as a result. Nothing about it requires the leading ant to know anything.

Agriculture, with crop protection

Leafcutter ants farm a fungus, and carry antibiotic-producing bacteria to protect the crop

Well supported

Good evidence backs this, though some details remain open.

Fungus-growing (attine) ants cultivate a fungal crop on cut plant material and host filamentous Pseudonocardia bacteria in specialised body structures; compounds from these bacteria suppress Escovopsis, a parasite specialised on attine fungus gardens.

Who this applies to
leafcutter ants and their fungus-growing relatives
Studied in
Atta, Acromyrmex

You may have heard

“Leafcutter ants eat leaves.”

They cannot digest leaves. The cut material is compost for a fungus, and the fungus is what they eat — along with a bacterial partner that keeps the crop healthy. Agriculture, monoculture and crop protection, tens of millions of years before anyone thought of them.

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

The farming itself is uncontroversial and long documented. The bacterial partnership was demonstrated by isolating the organisms and testing them against the specific parasite.

How far it can be extended

The association was documented across multiple attine genera, and fungus farming is the defining feature of the whole tribe.

Caveats

  • Antibiotic activity was shown in culture; quantifying protection inside a living colony is much harder.
  • How specific the bacterial partnership is has been debated in later work.

Still unanswered

  • How has the system avoided the resistance problems that defeat human antibiotics?

Last reviewed 2026-08-09

The evidence (1 study)

Leafcutter ants cannot digest leaves. The cut material is compost for a fungus, and the fungus is the crop.

Based on Leafcutter ants farm a fungus, and carry antibiotic-producing bacteria to protect the crop

Related

The mirror test claim

The claim that ants pass the mirror test is not supported

Popular claim, unsupported

Widely repeated, with no good evidence behind it.

The widely circulated assertion that ants demonstrate mirror self-recognition rests on a small study published in a venue with unclear peer review, using subjective scoring and no independent replication. No accepted evidence supports mirror self-recognition in any insect.

Who this applies to
the two ant species in the study behind the claimDo not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Myrmica sabuleti, Myrmica ruginodis

You may have heard

“Ants can recognise themselves in a mirror.”

It comes from one small study in a journal with an unclear review process, using subjective scoring, never independently reproduced. An ant touching a mark on its own head has explanations far simpler than self-recognition, starting with irritation.

Why we rate it this way, and what the caveats are
Popular claim, unsupportedHigh confidence

We are confident about the state of the evidence rather than about ant cognition: the only supporting study has serious methodological and publication problems, and nothing independent has reproduced it.

How far it can be extended

There is no adequate evidence for the claim in the species tested, so extending it to ants generally compounds a weak result with an unsupported leap.

Caveats

  • This entry exists because the claim circulates widely, not because there is a scientific question at issue.
  • Absence of good evidence is not proof of absence — but it is not evidence of presence either.

Still unanswered

  • Is the mirror test meaningful at all for an animal whose world is chemical rather than visual?

Last reviewed 2026-08-09

The evidence (1 study)

Full answer

Claims about this, checked

Things people have heard, and what the evidence actually supports.

The research behind this page

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

You may also find this interesting

What this page is still missing

NatureHQ publishes its own gaps. This record is at 97% completeness against what we would call a finished subject.

  • 2 high-priority search intent(s) not yet covered
  • Colony lifecycle, castes and reproduction are not yet covered.
  • Invasive ant species and their ecological effects would be a substantial addition.

Last reviewed 2026-08-09 · 8 claims · 76 search questions answered on this page