Skip to content
NatureHQ

Ecologyphenomenon

Extreme survival

An emperor penguin in a blizzard is about as warm as you are. Most animals in extreme places are not enduring the extreme — they are avoiding it, and that is a different story.

Most animals in extreme places are not enduring the extreme. They are escaping it — by finding a mild pocket of a hostile landscape, by shutting down until the season passes, or by holding a comfortable interior inside a hostile exterior. Genuine tolerance, where the tissue itself takes the punishment, is much rarer.

The interesting question about an animal in a hard place is not how tough it is, but whether it is experiencing the hardship at all. A great deal of what looks like endurance turns out, measured, to be evasion. An emperor penguin in a blizzard maintains a body temperature within a couple of degrees of a person’s; the extreme is outside it, held off by insulation and by a circulation that keeps the cold blood away from the warm core. A ground squirrel in an Alaskan winter is not enduring the winter either — it is absent from it, in a state where its metabolism runs at a small fraction of normal and months pass. Both are legitimately Arctic animals. Neither is tolerating the Arctic in the sense that the phrase implies. Real tolerance — a tissue that freezes solid and works again, a cell that dries to a crisp and rehydrates — exists, and it is worth separating out precisely because it is rarer and stranger than the avoidance strategies that surround it. Two other things follow. Tolerances are specific rather than general: the animals that survive freezing are frequently poor at surviving heat, and being tough at one thing predicts remarkably little about being tough at another. And every tolerance has terms attached — a state the animal has to be in beforehand, a duration beyond which it fails, and a fraction of individuals that die anyway. Those terms are not qualifications on the finding. They are the finding.

Developed coverage · 69% complete · reviewed 2026-09-03

What this page covers

Extreme tolerance occurs in every kingdom, but it is not evenly spread. The most extreme cases in animals are concentrated in small, aquatic or semi-aquatic invertebrates — tardigrades, rotifers, nematodes, brine shrimp — and the records for heat, acid and radiation belong to microbes rather than to animals at all.

Often confused with: Living somewhere extreme, which usually means avoiding the extreme rather than experiencing it; Being hard to kill in general — tolerances are specific, and an animal that survives freezing may die of mild warming; Extremophiles, which require their extreme conditions rather than merely surviving them

Quick facts

The commonest strategy
Avoidance — escaping the extreme in space, in time, or behind insulation
What a survival claim needs
A starting state, a duration, and a survival rate — all three, or it is not a finding
The word to be careful with
“Adapted” means three different things, on three different timescales
Toughness is not general
Tolerating one extreme predicts little about tolerating another

Avoiding it and enduring it are different solutions

The first question about any animal in a hard place is whether the hardship reaches it.

The short answer

Do animals in extreme places actually experience the extreme?

Usually not. Most survive by keeping the extreme away — inside a burrow, inside a season they sleep through, or inside a body that holds its own conditions. Genuine tissue-level tolerance is the minority strategy.

This is not a technicality; it changes what you should expect the animal to be able to do. An animal that avoids a condition is depending on the thing that lets it avoid — the snowpack, the burrow, the timing of a season — and if that disappears, the avoidance disappears with it, however tough the animal looks. An animal that tolerates the condition carries its solution in its tissues and takes it everywhere. Under a changing climate those two situations behave completely differently, which is why the distinction is worth insisting on even when both animals are, in ordinary speech, surviving the same winter.

Check it for yourself

Living somewhere extreme usually means not experiencing the extreme. Most animals in hard places escape the worst of it — in space, in time, or by holding a mild interior — rather than enduring it.

Well supported

Good evidence backs this, though some details remain open.

Survival in physically extreme environments is achieved predominantly through avoidance strategies — microhabitat selection, temporal escape into dormancy, and homeostatic regulation of the internal environment — rather than through tolerance of the extreme condition at the tissue level.

Who this applies to
Applies across animals in cold, hot, dry, low-oxygen and high-pressure environments; the balance between avoidance and tolerance differs by group.
Studied in
Animalia
Why we rate it this way, and what the caveats are
Well supportedModerate confidence

Individually well-evidenced across many systems, and unarguable for particular cases such as the Arctic ground squirrel. Stated as a generalisation about “most” it is a summary of many studies rather than a measured proportion, which is why the confidence is moderate rather than high.

How far it can be extended

The pattern is documented independently in polar, desert, alpine and deep-sea faunas, though the proportion of species relying on each strategy has not been quantified for any whole community.

Caveats

  • Avoidance and tolerance are not exclusive: many animals do both, and the same species may avoid a condition in one season and tolerate it in another.
  • “Most” here summarises a wide literature rather than reporting a count. No one has measured the proportion for a whole community.

Still unanswered

  • Whether avoidance strategies are more or less vulnerable to a changing climate than tolerance strategies — a microhabitat can disappear, while a tissue tolerance cannot.

Last reviewed 2026-09-03

The evidence (2 studies)
The same temperature, three different animals, three different situations
AnimalAt −3 °C, its tissue isStrategy
Emperor penguinAt about 38 °C, unaffectedInsulation and circulation: the cold never arrives
Arctic ground squirrelAt −3 °C and liquid, supercooledAvoidance of freezing while below freezing
Wood frogAt −3 °C and largely frozen solidTolerance: the ice is permitted, and controlled

The three rows are worth reading twice. Only the third animal is doing what the phrase “surviving extreme cold” suggests. The first is not cold. The second is cold and unfrozen, in a supercooled state that a single ice crystal would end. The third has allowed roughly two-thirds of the water in its body to freeze, which is the part that sounds impossible and is in fact the best-understood of the three.

Diagram

The same temperature, three different situations

Air at −3 °C, and what that means inside three animals.

Outside air at −3 °C, three animals, three situationsEmperor penguin≈ 38 °CNot cold at allInsulation, and heat kept inGround squirrel−3 °CCold, and unfrozenSupercooled — avoids iceWood frog−3 °CCold, and frozenTolerates ice, in controlOnly the third is doing what “surviving cold” suggests.The first two are avoiding the condition, by different means.
The same explanation in words

Three panels, all at an outside air temperature of −3 °C. An emperor penguin’s body is at about 38 °C: it is not cold at all, because insulation and its circulation keep the heat in. An Arctic ground squirrel’s body is at −3 °C and still liquid — supercooled, avoiding ice rather than tolerating it. A wood frog’s body is also at −3 °C and is frozen, tolerating the ice under control. Only the third is doing what the phrase “surviving extreme cold” suggests; the first two are avoiding the condition by different means.

Every survival has terms attached

A starting state, a duration, a survival rate. Remove any of the three and a finding becomes folklore.

No animal is immune to anything. Every demonstrated survival of an extreme came from a particular starting state, lasted a measured time, and killed some fraction of the animals tested.

Established

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

Reported tolerance of extreme conditions is conditional on physiological preparation, exposure duration and intensity, and is expressed as a survival rate rather than as an absolute capacity. Tolerance claims stated without these parameters are not interpretable.

Who this applies to
Applies to every reported case of extreme tolerance in animals.
Studied in
Animalia
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

This is a statement about the structure of the evidence rather than a hypothesis about organisms, and every study in this area exemplifies it.

How far it can be extended

Every experimental tolerance study reports conditions and a survival fraction; the pattern is a property of how the evidence is produced.

Caveats

  • This does not mean reported tolerances are exaggerated. They are frequently astonishing. It means the conditions are part of the finding rather than fine print.

Still unanswered

  • For most tolerant species, what the long-term cost of a survived extreme is: survival is scored at days, and lifetime consequences are rarely followed.

Last reviewed 2026-09-03

The evidence (3 studies)

The best-known sentence in this whole field is that tardigrades survive the vacuum of space. It comes from a real experiment, and almost every popular retelling removes the same three things. The animals were dried out first — a hydrated tardigrade in vacuum dies. The exposure was ten days. And the ones that also received unfiltered solar ultraviolet mostly did not come back; a handful of one species did. What was shown is remarkable enough without editing: an animal in a dried state survived ten days of vacuum, and sunlight, not vacuum, was the thing that killed it.

The full check on that claim

  • The starting state. Almost every extreme tolerance requires preparation — dried, cold-acclimated, fattened, or in a particular life stage. The same animal in the wrong state dies at conditions it would otherwise shrug off.
  • The duration. Tolerances are rated in hours or days far more often than indefinitely, and survival usually falls off sharply past some point rather than tapering.
  • The rate. “Survives” in a paper generally means a percentage, and often not a large one. Where a study reports it, this site reports it.
  • The cost. Survival is scored at the end of the experiment. What a survived extreme does to an animal’s later lifespan or breeding is rarely followed up, and is one of the real gaps in this field.

How we know

Taking space apart, to find out which part is lethal

Tardigrades survive exposure to space — but space delivers vacuum, cold and unfiltered sunlight at once. Which of them actually matters?

Desiccated tardigrades of two species were flown on an orbital mission and exposed for ten days in open space under separated conditions: vacuum alone, vacuum plus ultraviolet filtered to remove the shortest wavelengths, and vacuum plus the full solar ultraviolet spectrum. Animals were rehydrated afterwards and survival scored.

What happened

Survival of vacuum alone was high and close to the ground controls. Adding the full solar ultraviolet spectrum reduced survival drastically, with only a small proportion of one species recovering. Some survivors went on to reproduce.

What it shows

That vacuum — the factor the popular version of this result is entirely about — is very nearly survivable for an animal already in its dried state, and that ultraviolet radiation is what kills. Separating the conditions is what converts "they survived space" into something you can learn from.

What it does not show

It does not show that a living, hydrated tardigrade survives vacuum; that was not tested and could not be, since an active tardigrade in vacuum dies at once. It does not extend beyond ten days. And low Earth orbit is not deep space, whose radiation environment is substantially harsher — which is why this result argues against the panspermia use often made of it rather than for it.

The controls — what makes this evidence rather than a story
  • Ground controls held in the same desiccated state for the same ten days, so that the cost of desiccation itself is separated from the cost of exposure.
  • The three exposure conditions differ in one factor at a time, which is the whole point of the design.
  • Two species flown, so that a result peculiar to one animal would be visible as such.

From Tardigrades survive exposure to space in low Earth orbit

Adapted, acclimatised, tolerant

Three words that get used interchangeably and mean three different timescales.

An individual animal cannot adapt. It can acclimatise — remodel itself over days or weeks, reversibly. Adaptation is a change in a population across generations, and no individual undergoes it.

Established

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

Adaptation denotes a change in the genetic composition of a population across generations under selection; acclimatisation denotes reversible phenotypic adjustment within an individual’s lifetime. The two differ in timescale, reversibility, level of organisation and mechanism.

Who this applies to
A distinction that applies wherever organisms respond to conditions — the terms mean the same thing in plants, animals and microbes.
Studied in
Animalia, Plantae
Why we rate it this way, and what the caveats are
EstablishedHigh confidence

Definitional and universally accepted within the field. What is contested is how much of each occurs in any given case, not whether they differ.

How far it can be extended

The distinction is definitional in evolutionary biology and physiology rather than a finding about any particular group, and it is applied identically across taxa.

Caveats

  • The boundary is not always clean. Developmental plasticity — an adjustment made once, early, and not reversed — sits between the two, and epigenetic effects that persist for a generation or two complicate it further.
  • Ordinary usage will not change, and “adapted” in general speech usually means something closer to “coped”. The distinction matters when the question is whether a species can keep coping.

Still unanswered

  • How much of the variation in tolerance seen between populations of one species is genetic and how much is a lasting effect of the conditions individuals developed in.
  • Whether high acclimatisation capacity slows genetic adaptation by hiding variation from selection.

Last reviewed 2026-09-03

The evidence (3 studies)

The distinction in full

  • Adaptation

    What the word means strictly, and why the strict meaning matters

The strategies, and where to read about each

Extreme survival is not one subject. It is several, grouped by what is being survived.

Shutting down until it passes

  • Dormancy

    The shared machinery behind all of the states below

  • Torpor

    Hours to days, and measurably distinct from hibernation

  • Hibernation

    Months, and not a long sleep

  • Diapause

    Programmed in advance, sometimes a generation ahead

  • Cryptobiosis

    Where metabolism cannot be detected at all

Taking the conditions on directly

What is not settled

  • Does surviving an extreme shorten an animal’s life afterwards?

    Why it matters: Tolerance studies score survival at the end of the exposure, typically within days. If a survived freezing or drying costs an animal months of later life or a breeding season, the ecological meaning of “survived” changes considerably.

    What would settle it: Long-term follow-up of individually marked animals after a measured extreme exposure, compared against unexposed controls of the same cohort.

  • Are avoidance strategies more fragile than tolerance strategies under climate change?

    Why it matters: An animal that depends on a snowpack, a burrow microclimate or a reliable seasonal cue loses its strategy when those change, while a tissue-level tolerance travels with the animal. If avoidance dominates, as it appears to, the vulnerability of cold-adapted faunas may be systematically underestimated.

    What would settle it: Community-level surveys classifying species by strategy and tracking their fates through documented climatic shifts.

  • Why does tolerance of one extreme predict so little about tolerance of another?

    Why it matters: Popular accounts treat toughness as a single quality. The evidence suggests largely independent mechanisms, which would mean there is no such thing as a generally indestructible animal — only animals with specific, unrelated capacities.

    What would settle it: Systematic multi-stressor testing across many species in one laboratory, which is rarely done because each stressor is a separate specialism.

The research behind this page

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

2020Cell Communication and Signaling

The biology of tardigrade disordered proteins in extreme stress tolerance

Several distinct protein families contribute, they are not interchangeable, and no single one accounts for tardigrade tolerance; the mechanisms are only partly understood.

2019Nature Ecology & Evolution

Morphology and genome of a snailfish from the Mariana Trench

The fish has an incompletely ossified skull, unusually flexible bones, no swim bladder, and changes in genes governing bone development and membrane composition.

2013Journal of Experimental Biology

Avoidance and tolerance of freezing in ectothermic vertebrates

Avoidance and tolerance are alternative solutions with different failure modes, and the strategy a species uses tracks where it overwinters rather than how cold it gets.

2011Nature

Climate change and evolutionary adaptation

Documented genetic adaptation to recent warming is far rarer than documented phenotypic response, and much of the observed change in wild populations is plasticity rather than evolution.

2010Journal of Experimental Biology

The physiology of climate change: how potentials for acclimatization and genetic adaptation will determine “winners” and “losers”

Acclimatisation capacity and adaptive capacity are different quantities, they are not correlated across species, and a species can have a wide tolerance range while having almost no ability to shift it.

2008Current Biology

Tardigrades survive exposure to space in low Earth orbit

Survival of vacuum alone was high, comparable to ground controls.

2008Proceedings of the National Academy of Sciences

Impacts of climate warming on terrestrial ectotherms across latitude

Tropical species live closest to their thermal optima and have the narrowest safety margins, so modest warming reduces their performance while temperate species, living well below their optima, may initially benefit.

2007Annual Review of Physiology

Hypoxia tolerance in reptiles, amphibians, and fishes: life with variable oxygen availability

Anoxia-tolerant vertebrates suppress metabolism dramatically, reduce the electrical activity of the brain in a controlled way rather than losing it, and manage the resulting acid — in turtles by dissolving buffer from their own shells.

2006Journal of Experimental Biology

When bad things happen to good fish: the loss of hemoglobin and myoglobin expression in Antarctic icefishes

The evidence indicates loss without compensating advantage: icefish sustain oxygen delivery through very large hearts, high blood volume and low metabolic demand, at considerable circulatory cost.

2004Journal of Experimental Biology

Hypoxic survival strategies in two fishes: extreme anoxia tolerance in the North European crucian carp and natural hypoxic preconditioning in a coral-reef shark

Crucian carp survive months of complete anoxia in ice-covered ponds by converting lactate to ethanol and excreting it across the gills, avoiding the acid build-up that kills other fish.

2003Science

Acclimation capacity underlies susceptibility to climate change

The species with the highest thermal limits had the least capacity to raise them further.

2002Science

Unpredictable evolution in a 30-year study of Darwin’s finches

Beak dimensions shifted measurably within a few generations in response to changes in the available seeds, and shifted back and sideways as conditions changed.

2001Comparative Biochemistry and Physiology Part B

Cryptobiosis — a peculiar state of biological organization

In the deepest cryptobiotic states no metabolism can be measured by available methods, and the organism’s survival depends on the physical preservation of its structures rather than on any ongoing repair.

2001Nature

Life in extreme environments

The recorded limits of life are set overwhelmingly by microbes rather than by animals, and most extremophiles require their conditions rather than merely tolerating them — an acidophile grown at neutral pH dies.

1999Biological Reviews

Metabolic depression in animals: physiological perspectives and biochemical generalizations

Depressed states across very distant phyla converge on a small set of mechanisms — coordinated suppression of protein synthesis and ion pumping, reversible phosphorylation of enzymes, and a general shutdown rather than a targeted one — despite having been named and studied independently.

1998Annual Review of Physiology

Biological ice nucleation and ice distribution in cold-hardy ectothermic animals

Freeze-tolerant animals often produce ice-nucleating proteins that trigger freezing at high sub-zero temperatures in extracellular spaces, while freeze-avoiding animals do the opposite — clearing their guts and removing nucleators to stay liquid as far below zero as possible.

1992Annual Review of Physiology

Adaptations to high hydrostatic pressure

Pressure destabilises proteins and stiffens membranes because both processes involve volume changes.

1989Science

Freeze avoidance in a mammal: body temperatures below 0 °C in an Arctic hibernator

Body temperature fell below freezing — to around −3 °C — and was held there for weeks without the animal freezing, in a supercooled state, while the squirrel remained alive and periodically rewarmed.

1988Physiological Reviews

Freeze tolerance in animals

Freeze-tolerant animals permit ice in extracellular spaces while keeping cells themselves unfrozen, accumulate low-molecular-weight cryoprotectants such as glucose and glycerol, and frequently use proteins that deliberately start ice formation in safe locations at high sub-zero temperatures.

1981Science

Prokaryotic cells in the hydrothermal vent tube worm Riftia pachyptila: possible chemoautotrophic symbionts

The worm houses chemoautotrophic bacteria that oxidise hydrogen sulphide to fix carbon, and is nourished by them.

1979Proceedings of the Royal Society of London B

The spandrels of San Marco and the Panglossian paradigm: a critique of the adaptationist programme

A trait can exist because of developmental constraint, because it is a by-product of something else, because of genetic drift, or because it was built for a different purpose.

1964Oxford University Press

Desert Animals: Physiological Problems of Heat and Water

Desert animals survive by budgeting rather than by endurance.

1957American Journal of Physiology

Body temperature of the camel and its relation to water economy

A dehydrated camel allows its body temperature to swing by around six degrees across the day — cooling overnight and rising through the morning without sweating — and re-radiates the stored heat at night.

1954Nature

Vertebrates without erythrocytes and blood pigment

These fish have essentially no red blood cells and no haemoglobin.

1952Physiological Reviews

Water metabolism of desert mammals

Kangaroo rats survive indefinitely without drinking, on water produced by metabolising their food, because losses are cut on every route at once: extremely concentrated urine, nearly dry faeces, no sweating, and recovery of water from exhaled air.

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 69% completeness against what we would call a finished subject, and was last reviewed on 2026-09-03. It carries 12 claims and answers 2 mapped search questions.

  • no popular claim about this subject has been checked yet
  • The heat, pressure and low-oxygen strategies are named here but their detailed pages are not yet written.
  • Microbial extremophiles are mentioned as the holders of most records and are not covered in their own right.
  • The long-term cost of a survived extreme is an open question across the whole family rather than a gap in one page.