Hibernation is metabolism turned down, not sleep prolonged. A deep hibernator runs at a few per cent of its normal rate for weeks at a time — and the suppression comes first, with the drop in body temperature following it rather than causing it.
The intuitive picture of hibernation has the causation backwards. It looks as though the animal gets cold and therefore goes slow, physics doing the work. What the measurements show is the opposite: metabolic rate is actively suppressed first, and body temperature falls as a consequence. That single fact reorganises the whole subject. It explains why a bear, which cannot cool to near-freezing because a body that large simply will not, can still be hibernating in every sense that matters — the objection to bears was always an objection about temperature, and temperature turned out to be the wrong variable. It explains why hibernation is not a passive state you slip into but one an animal enters and controls. And it sets up the strangest finding in the area: hibernators do not stay under. Every few days to weeks they rewarm to normal temperature for several hours, at a cost of a substantial share of the fat that has to last all winter, and what they appear to do with that expensive time is sleep. Whatever sleep is for, a week of torpor does not supply it.
Developed record · 69% complete · reviewed 2026-08-10
What this page covers
A physiological state rather than a group of organisms. Best characterised in small mammals, with bears the case that forced the definition to be examined.
Often confused with: Sleep, which is a different state with different brain activity; Aestivation, the equivalent response to heat and drought rather than cold; Brumation, the term used for the winter dormancy of reptiles
Quick facts
What it is
Actively suppressed metabolism; temperature follows
Not sleep
Hibernators arouse periodically and sleep then
Bears
Metabolism to ~25% of basal; temperature only to 30–36°C
Deep hibernators
A few per cent of basal metabolic rate, body temperature near ambient
A hibernating animal turns its metabolism down first, and gets cold as a consequence
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Metabolic rate in hibernating mammals falls to as little as a few per cent of basal rate, and the suppression begins before body temperature declines rather than following from it. Hibernation and daily torpor differ quantitatively in bout length, minimum body temperature and depth of suppression, forming a continuum of controlled heterothermy rather than two unrelated states.
Who this applies to
heterothermic mammals, with comparative data dominated by small laboratory hibernators
Studied in
Mammalia, Chiroptera, Sciuridae
You may have heard
“Hibernating animals sleep through the winter”
Two errors. It is not sleep — hibernators appear to arouse periodically in order to sleep, at enormous energetic cost. And the animal is not slow because it is cold; it turns its metabolism down and the temperature follows, which makes hibernation a state entered rather than one succumbed to.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Directly measured by simultaneous metabolic and temperature recording, where the order of events distinguishes active suppression from passive cooling.
How far it can be extended
Active metabolic suppression preceding cooling has been measured across multiple hibernating species and in daily torpor.
Caveats
Comparative data are dominated by small mammals that hibernate conveniently in laboratories.
Field metabolic rates are scarcer than laboratory ones and tend to be higher.
Where to draw the line between deep daily torpor and shallow hibernation is a matter of convention.
Still unanswered
What triggers the periodic arousals that consume much of a hibernator’s fat budget?
Metabolic suppression to roughly a quarter of basal rate with only a modest temperature fall — suppression independent of cooling.
A hibernating arctic ground squirrel can hold a body temperature below zero, breathe a few times a minute, and run at a small fraction of the energy it uses in summer. It is tempting to explain that thermally — cold chemistry runs slowly, so a cold animal is a slow one. The recordings do not support that reading. Metabolic suppression begins before body temperature starts to fall, which means the animal is turning something down rather than cooling into a stupor.
Metabolic rate: to as little as 2–5% of basal in deep hibernators.
Body temperature: often within a degree or two of the surrounding air, and below freezing in a few species.
Heart rate: from hundreds of beats per minute to single figures, with long pauses.
Breathing: minutes may pass between breaths.
Fuel: fat laid down in advance, plus in some species a cache of food eaten during arousals.
Brown adipose tissue does the rewarming. It is fat specialised for producing heat rather than storing energy, concentrated between the shoulder blades in many hibernators, and it is what allows an animal at near-freezing to raise its own temperature by thirty degrees in a couple of hours without shivering its way there.
The suppression is not all-or-nothing over the season. Hibernation is a sequence of torpor bouts separated by arousals, and the bouts get longer as winter deepens and shorter again towards spring.
Words used here
Torpor bout
One continuous period of suppressed metabolism, from entry to the next arousal. Days to weeks in a deep hibernator.
Brown adipose tissue
Fat specialised for generating heat rather than storing energy. The rewarming engine of a hibernator.
Heterothermy
Allowing body temperature to vary substantially rather than holding it constant. The general category torpor and hibernation sit in.
These terms describe points along a continuum, and the field does not draw the boundaries in exactly the same place. That is worth saying plainly rather than presenting a tidy taxonomy that different papers would contradict.
The states, and what separates them
State
Duration
Typical depth
Example
Daily torpor
Hours, within a day
Metabolism to roughly 10–30% of basal
Hummingbirds overnight; many small bats and mice
Hibernation
Bouts of days to weeks, across a season
To a few per cent of basal; temperature near ambient
Ground squirrels, dormice, many temperate bats
Winter dormancy in bears
Months, with no full arousal
To ~25% of basal; temperature only to 30–36°C
Black and brown bears
Aestivation
Weeks to months
Varies; response to heat and drought rather than cold
Lungfish, some snails, a few mammals
Brumation
A season
Ectotherm; metabolism falls with ambient temperature
Reptiles and amphibians overwintering
The last row is the one that most often causes confusion, and the distinction is real. A hibernating mammal is suppressing its own metabolism against a temperature it could otherwise maintain. An overwintering lizard is simply cold, with no thermostat to override — which is why "brumation" exists as a separate word rather than being folded in.
Some hummingbirds enter torpor most nights, dropping body temperature by twenty degrees or more, and resume feeding within minutes of dawn. Torpor is not rare or exotic; for a very small endotherm it is routine housekeeping.
Words used here
Aestivation
Dormancy in response to heat or drought rather than cold.
Brumation
Winter dormancy in reptiles and amphibians. Distinguished from hibernation because an ectotherm cools passively.
The evidence is that hibernators interrupt it in order to sleep.
Hibernation is not sleep — hibernators spend fat rewarming so that they can sleep
Emerging evidence
Real findings exist, but too few or too recent to be settled.
Hibernation is interrupted every few days to weeks by periodic arousals during which body temperature returns to normal for several hours at very large energetic cost. Electroencephalographic recording during these arousals in ground squirrels shows slow-wave sleep, and the amount recorded is greatest following the longest torpor bouts — the pattern expected if sleep pressure accumulates during torpor rather than being discharged by it.
Who this applies to
ground squirrels recorded in the laboratory through hibernation and arousalDo not extend this beyond the taxa listed — the popular version over-reaches.
Studied in
Spermophilus
You may have heard
“Hibernation is a really deep sleep”
It is a different state that looks similar from outside. The strongest evidence against the equation is that hibernators interrupt it — rewarming every week or two at a cost of a substantial fraction of their winter fat, and spending that expensive time asleep. Whatever sleep provides, a week of torpor does not appear to provide it.
Why we rate it this way, and what the caveats are
Emerging evidenceModerate confidence
A striking and well-designed measurement with the predicted dose-response between torpor duration and subsequent sleep. Confidence is moderate because it is one genus, interpreting an electroencephalogram at low body temperature is contested, and correlation does not establish that sleep pressure causes the arousal.
How far it can be extended
Demonstrated in one genus. The interpretation is influential and not universally accepted, and other functions for arousals have been proposed.
Caveats
One genus, in the laboratory, with implanted electrodes.
Interpreting the electroencephalogram at low body temperatures is contested.
Other functions for periodic arousals — immune, excretory, metabolic — are also proposed and not excluded.
Still unanswered
Do all hibernators show the same sleep pattern during arousals, or is this specific to ground squirrels?
If sleep pressure drives arousal, what accumulates during torpor to create it?
The criteria that make sleep and torpor distinguishable states rather than one continuum.
Periodic arousals are the strangest feature of hibernation and the hardest to explain away. Every few days to weeks a hibernator rewarms all the way to normal body temperature, stays there for several hours, and cools again. It is enormously expensive: across a winter, arousals can consume the majority of the energy the animal spends, in an animal whose entire strategy is to spend as little as possible.
How we know
Recording a hibernator during the hours it wakes up
Hibernators rewarm every week or two at enormous energetic cost, then cool again. What is worth that much of a winter fat budget?
Rather than recording torpor, this recorded the interruptions. Ground squirrels with implanted electrodes were monitored through hibernation, and brain activity was scored during the periodic arousals when body temperature returns to normal for a few hours. The critical measurement is relational: the amount of slow-wave sleep during each arousal was compared against the length of the torpor bout that preceded it.
What happened
Arousals showed the electroencephalographic signature of slow-wave sleep, and the amount recorded was greatest following the longest torpor bouts.
What it shows
That torpor and sleep are different states, and — if the dose-response is causal — that hibernation accrues sleep pressure rather than discharging it. That would make the most expensive component of hibernation a debt being repaid, which is a strong claim about what sleep is for: something a week of cold immobility does not provide.
What it does not show
A correlation between bout length and subsequent sleep does not establish that sleep pressure causes the arousal; several other functions have been proposed for arousals — immune activity, waste clearance, water balance — and none is excluded. One genus, in the laboratory, with implanted electrodes. Interpreting sleep architecture in an animal that has just rewarmed from near-freezing is not straightforward.
The controls — what makes this evidence rather than a story
Torpor bout length recorded for every interval, so the sleep measurement can be tested against a graded predictor rather than a single condition.
Recording only at normal body temperature during arousals, avoiding the contested question of how to interpret an electroencephalogram in a near-frozen brain.
Continuous monitoring across many bouts per animal, making each animal its own comparison.
Body temperature logged throughout, so arousal onset and completion are defined by measurement rather than by observation.
If the finding holds, hibernation is not merely different from sleep — it is a state during which sleep cannot happen and pressure for it builds. That is a strong claim from one genus, and NatureHQ carries it as early evidence rather than as settled. Other functions for arousals are proposed too, including immune activity, clearing metabolic waste and restoring water balance, and they are not mutually exclusive.
Bears do hibernate — the old objection measured the wrong variable
Established
Specialists would state this without hedging. Multiple independent lines of evidence agree.
Black bears in winter dens reduce metabolic rate to approximately 25% of basal rate while body temperature declines only to about 30–36°C, in slow multi-day cycles, with metabolic suppression persisting for weeks after body temperature normalises in spring. Because the suppression is largely independent of temperature, the modest thermal decline that formerly excluded bears from definitions of hibernation does not indicate a modest metabolic response.
Who this applies to
American black bears measured through a full winter season
Studied in
Ursus americanus
You may have heard
“Bears do not really hibernate”
A reasonable position while the only measured variable was body temperature, which in a bear falls a few degrees rather than to near freezing. Measuring metabolism directly changed the answer: suppression to roughly a quarter of basal rate, sustained for months. A body the size of a bear cannot cool like a squirrel — the physics forbid it — so temperature was always the wrong thing to define the state by.
Why we rate it this way, and what the caveats are
EstablishedHigh confidence
Continuous metabolic, temperature, heart rate and activity recording through a full season in undisturbed dens — the measurement whose absence sustained the earlier disagreement.
How far it can be extended
Measured directly in black bears; brown bears show a broadly comparable pattern, and polar bear denning differs in season and circumstance.
Caveats
Five animals in artificial dens, however undisturbed the site.
Whether bears "count" as hibernators still depends on which definition is used; this settles the physiology, not the terminology.
A denning bear can rouse and defend itself, which a deeply torpid ground squirrel cannot — a real difference the metabolic definition does not capture.
Still unanswered
How do bears avoid the muscle and bone loss that immobility produces in other mammals?
The comparative framework within which a large mammal cannot cool as far as a small one regardless of how suppressed it is.
For decades the textbook answer was no, and the reasoning was not silly. Hibernation was defined by how far body temperature fell, a hibernating ground squirrel drops to near freezing, and a denning bear drops only a few degrees. On that definition bears were doing something else, and the term "winter dormancy" was coined for it.
How we know
Measuring a bear all winter instead of arguing about it
A denning bear’s body temperature falls only a few degrees, where a hibernating ground squirrel approaches freezing. Does that mean bears are not really hibernating?
The disagreement had persisted because the only variable routinely measured in a denning bear was body temperature, and temperature had been made the definition. This measured the thing the definition was standing in for. Black bears were housed in artificial dens in undisturbed forest and instrumented to record oxygen consumption, body temperature, heart rate and muscle activity continuously across a full winter, with the site left alone so that the animals were not repeatedly roused by the act of measurement.
What happened
Metabolic rate fell to roughly 25% of basal while body temperature declined only to about 30–36°C in slow multi-day cycles. Heart rate dropped dramatically with long pauses between beats, and metabolic suppression persisted for several weeks after body temperature had returned to normal in spring.
What it shows
That metabolic suppression in bears is largely independent of body temperature, so a modest thermal decline is not evidence of a modest response. It also explains why the disagreement lasted: a body the size of a bear cannot cool like a squirrel whatever its metabolism does, so temperature was never a fair test of the state.
What it does not show
It does not settle the terminology, only the physiology — whether bears "count" still depends on which definition is adopted, and there remains a real difference the metabolic definition does not capture, since a denning bear can rouse and defend itself where a deeply torpid ground squirrel cannot. Five animals of one species, in artificial dens however undisturbed, and prior handling affects an animal whose entire response is to being left alone.
The controls — what makes this evidence rather than a story
Metabolic rate measured directly by oxygen consumption rather than inferred from body temperature, which is the entire point of the design.
Continuous recording across the whole season, so that multi-day cycles are observed rather than sampled.
Summer measurements from the same animals establishing each bear’s own basal rate, so suppression is a within-animal comparison.
Undisturbed dens in forest, since a bear roused by observers is not hibernating during the observation.
Recording continued into spring, which is what revealed suppression persisting after rewarming.
Measuring metabolism directly reversed the answer. A denning black bear runs at roughly a quarter of its basal rate for months, and the suppression persists for weeks after body temperature has returned to normal in spring. A body the size of a bear cannot cool to near freezing — the surface-to-volume ratio forbids it — so temperature was never going to be a fair test.
There is still a real difference worth keeping, and it is not about temperature. A deeply torpid ground squirrel cannot be woken quickly and is effectively defenceless; a denning bear can rouse, move and defend itself, and females give birth and nurse cubs during the winter. Calling both hibernation is defensible and does flatten something.
A denning bear does not eat, drink, urinate or defecate for months, recycling urea into protein. It emerges having lost fat but remarkably little muscle or bone, which is why the physiology is of interest to medicine.
Hibernation is scattered across the mammals rather than concentrated in one group, which suggests it has arisen repeatedly rather than being inherited from a common hibernating ancestor.
Rodents: ground squirrels, marmots, dormice — the classic deep hibernators, and most of the laboratory data.
Bats: many temperate species hibernate for months, and disturbance is a serious conservation issue because rewarming costs fat that cannot be replaced.
Hedgehogs and some insectivores.
Bears, by the metabolic definition.
One bird, the common poorwill, which is the only bird known to hibernate rather than use daily torpor.
And the animals commonly assumed to hibernate and mostly not: squirrels of the tree-dwelling kind stay active all winter on cached food; badgers and raccoons enter torpor in cold spells but not a season-long state; and insects overwinter through diapause, which is a developmental arrest rather than metabolic suppression in the mammalian sense.
The arc here is a definition following the instruments. While body temperature was the only thing that could be measured continuously in a wild animal, hibernation was defined by how cold something got — and every later correction came from measuring something else.
1832
First observation
Hibernation described as a physiological state
Marshall Hall and contemporaries establish that hibernating mammals show profoundly reduced respiration and heart rate rather than simply being asleep — an observation that then waits a century for the means to quantify it.
1960
Reinterpretation
Body temperature becomes the defining criterion
As temperature telemetry becomes routine, hibernation is characterised by how far body temperature falls, and animals whose temperature falls only modestly — bears among them — are placed in a separate category of winter dormancy.
Changes how the 1832 result reads
A reasonable definition built on the variable that could actually be measured continuously in a wild animal, and one that quietly assumed temperature and metabolism move together.
1991
Modern discovery
Hibernators are found to sleep during arousals
Ground squirrels recorded through periodic arousals show slow-wave sleep, in amounts scaling with the length of the preceding torpor bout — suggesting torpor accrues sleep pressure rather than discharging it.
Comparative work across torpor and hibernation establishes that metabolic suppression begins before body temperature falls, so the state is actively entered rather than passively fallen into.
Changes how the 1960 result reads
If metabolism leads and temperature follows, then defining hibernation by temperature measures the consequence instead of the state.
Continuous metabolic recording in denning black bears finds suppression to roughly a quarter of basal rate with body temperature falling only to 30–36°C, and persisting for weeks after spring rewarming.
Changes how the 1960 result reads
The exclusion of bears rested entirely on the temperature criterion. Measure the metabolism and the exclusion goes with it — the disagreement was about instruments, not about bears.
Why it matters: Arousals consume most of the energy hibernation is designed to save, so whatever they accomplish must be worth that price. Sleep, immune function, waste clearance and water balance are all candidates and none is established.
How do hibernators avoid the damage immobility causes?
Why it matters: Months of inactivity should cost muscle and bone, and largely does not. The mechanism is of direct medical interest and is not understood.
What sets the length of a torpor bout?
Why it matters: Bouts lengthen through the winter and shorten towards spring in a pattern too regular to be incidental, and what is being tracked is unknown.
How will hibernators fare as winters shorten?
Why it matters: Entry and emergence are cued partly by conditions that are shifting, and an animal that emerges before its food does has spent its fat for nothing.