Animal Sleep: How Different Species Sleep (And Why Some Barely Sleep at All)

For years, scientists assumed that sleep was a vulnerability, an unavoidable tax that evolution would minimize wherever possible. Yet a giraffe in the wild sleeps as little as 30 minutes a day while a brown bat sleeps 20 hours. A great frigatebird sleeps while flying across the open ocean. A newborn orca calf does not sleep at all for its first month of life, and neither does its mother. Sleep is not a single, uniform state that all animals share. It is an extraordinarily diverse set of biological solutions, shaped by predation pressure, body size, diet, and neural architecture that differs radically from one species to the next. The real mystery is not just why animals sleep, but why sleep looks so different everywhere biologists look for it.

Educational content, not medical advice. This article covers comparative animal physiology and behavior. It is not a substitute for veterinary or medical guidance about sleep in humans or pets.

Key takeaways

• Dolphins and many seabirds sleep with one hemisphere of the brain at a time, a state called unihemispheric slow-wave sleep, so they never lose full awareness of their surroundings.
• Wild African elephants sleep as little as 2 hours per night, the least documented sleep duration of any land mammal studied so far, while their captive counterparts sleep 4 to 6 hours.
• The great frigatebird can sleep mid-flight, confirmed by EEG recordings on free-flying birds crossing the Pacific Ocean.
• Newborn orcas and their mothers appear not to sleep at all for roughly the first month after birth, with no measurable rebound in sleep once normal patterns resume.
• REM-like sleep has now been documented in octopuses, an invertebrate lineage that split from vertebrates over 500 million years ago, suggesting either an ancient shared origin or a striking case of convergent evolution.

What counts as sleep, and why it is hard to study across species

Sleep is not defined by closed eyes or stillness. Biologists use a behavioral checklist that applies whether the animal is a mouse, a fruit fly, or a jellyfish-like organism with no brain at all. An animal is considered asleep when it shows sustained inactivity in a stereotyped posture, a higher threshold for waking than during normal rest, and a homeostatic rebound, meaning it sleeps more or more deeply after being deprived of sleep to make up the deficit.

That last criterion is the one that separates sleep from simply resting or being tired. A lion resting under a tree is not necessarily asleep in the biological sense. A lion that has been kept awake artificially and then sleeps unusually hard the next day, recovering lost sleep, is displaying the homeostatic signature that defines true sleep. This behavioral definition matters because most wild animals cannot be hooked up to an electroencephalogram (EEG), the standard tool for measuring brain-wave sleep stages in a lab. Researchers studying free-ranging elephants, whales, or birds have to infer sleep from movement sensors, accelerometers, and rare opportunities to attach lightweight EEG loggers, which is part of why animal sleep science has exploded only in the last two decades as wearable sensor technology improved.

The other complication is that "sleep" in comparative biology covers at least two distinct brain states in vertebrates: slow-wave sleep (deep, restorative, associated with large synchronized brain waves) and REM sleep (rapid eye movement sleep, associated with dreaming, muscle paralysis, and memory processing in mammals and birds). Different species allocate wildly different proportions of their sleep time to each state, and some animals, like dolphins, experience only fragments of REM-like activity at all.

Newborn orca calves and their mothers go without measurable sleep for roughly the first month of the calf's life, according to UCLA neuroscientist Jerome Siegel's research team, and neither shows a sleep rebound once normal rest resumes.

Unihemispheric sleep: how dolphins and birds stay half-awake

Bottlenose dolphins face a problem no land mammal has: breathing is a conscious act for them, not an automatic reflex. If a dolphin lost full consciousness the way a sleeping dog does, it would stop surfacing to breathe and drown. Evolution's answer is unihemispheric slow-wave sleep (USWS), in which one half of the brain enters deep slow-wave sleep while the other half stays awake, monitoring the water, controlling breathing, and watching for predators.

Think of it like a night-shift rotation at a factory that can never fully shut down. One team clocks out and rests while the other keeps the machinery running, and then they swap. Dolphins swim slowly near the surface during USWS, with one eye closed, connected to the resting hemisphere, and the other eye open, connected to the alert one. After a period, the hemispheres switch roles, and this can cycle for hours. Dolphins never appear to lose consciousness entirely, which for decades fed the myth that they do not sleep at all. They sleep constantly; it is simply distributed across the brain rather than experienced as one continuous blackout.

Many seabirds show a similar pattern, and the most dramatic documented case comes from a 2016 study in Nature Communications by Niels Rattenborg and colleagues at the Max Planck Institute for Ornithology. The team attached lightweight EEG sensors to great frigatebirds during multi-day nonstop flights over the Pacific Ocean and recorded clear episodes of unihemispheric sleep while the birds were airborne, gliding on rising air currents to conserve energy. Occasionally the birds entered brief bihemispheric sleep, both hemispheres resting simultaneously for a few seconds while gliding, a kind of microsleep snatched mid-flight. Across the whole migratory flight, the birds slept an average of only about 42 minutes per day, compared with roughly 12 hours per day when they were on land incubating eggs or resting on the water. That is a nearly 17-fold drop in sleep time, sustained for days at a stretch, without obvious cognitive breakdown during the flight.

Do any animals sleep with their whole brain shut off while still moving?

No confirmed case exists of a fully unconscious animal actively swimming or flying under its own navigational control. Unihemispheric sleep and brief bihemispheric microsleeps during flight or swimming are the closest analogues, allowing partial rest without total loss of motor control or environmental awareness. Resting posture is another curious animal behavior, explored in why flamingos stand on one leg.

The elephant paradox: why wild elephants sleep so much less than captive ones

Captive African elephants in zoos and sanctuaries typically sleep about 4 to 6 hours per day, based on multiple observational studies. Wild African elephants tell a very different story. In a 2017 study published in PLOS ONE, neuroscientist Paul Manger and colleagues fitted GPS collars and activity-logging accelerometers to two free-ranging matriarch elephants in Botswana's Chobe National Park and tracked them continuously. The wild elephants slept an average of just 2 hours per day, the shortest confirmed sleep duration of any land mammal studied to date.

The wild elephants also spent only about 8% of their time lying down, and true deep sleep with REM appears to require recumbency in elephants because standing sleep cannot include the muscle atonia (loss of muscle tone) that defines REM. On some days the study animals went entirely without lying down, sometimes for 48 hours or more, and covered up to 30 kilometers (about 19 miles) walking and foraging through the night. Manger's team proposes that the combination of predation risk, primarily from lions that target calves, and the sheer caloric demand of an elephant's diet, which requires up to 18 hours of feeding per day to meet metabolic needs, compresses sleep down to a bare minimum in the wild. The matriarch, who bears primary responsibility for monitoring the herd's safety, appeared to sleep the least of any individual tracked.

This finding matters beyond elephant biology because it challenges assumptions about how much sleep large-brained mammals actually require for cognitive maintenance. Current theories of sleep function center on memory consolidation, synaptic pruning, and clearance of metabolic waste products from brain tissue through the glymphatic system, a topic covered in depth in how sleep affects the brain. If two hours of fragmented sleep per night is sufficient to sustain the complex social cognition, long-term memory, and problem-solving that elephants are known for, then either sleep is dramatically more efficient in some neural architectures than researchers assumed, or land mammals have back-up mechanisms for some of sleep's restorative functions that remain unidentified. Neither explanation is fully settled, and Manger's team has explicitly flagged the finding as a challenge to existing sleep theory rather than a solved case.

Giraffes: the fragmented, 9-minute-nap survival strategy

Giraffes push the elephant pattern even further. Classic behavioral observations of wild giraffes found total sleep time ranging from roughly 30 minutes to 2 hours per 24-hour period, and some field studies recorded average deep-sleep bouts of only a few minutes at a time, broken into a dozen or more micro-naps scattered across the night. A giraffe's vulnerability while lying down, its long legs make standing back up slow and awkward, and a recumbent giraffe is an easy target for lions, means true recumbent sleep with REM is rationed carefully and almost always shared among a herd so at least one animal remains alert.

Captive giraffes, freed from predation risk, behave very differently. An early study of eight giraffes in a zoological garden, published in the South African Journal of Wildlife Research, recorded an average of about 4.6 hours of sleep per day, with animals far more willing to lie down for extended stretches. The gap between wild and captive sleep in giraffes mirrors the elephant pattern almost exactly: strip away predation risk and energetic pressure, and sleep duration climbs several-fold. That consistency across two unrelated large herbivores is itself informative. It suggests sleep duration in large-bodied prey mammals is not fixed by biology alone but is actively suppressed by environmental risk, and the "true" biological requirement may sit somewhere between the wild and captive extremes.

Horses: standing sleep, and why REM still requires lying down

Horses split the difference between elephants and dolphins in an interesting way. Adult horses need roughly 5 to 7 hours of total rest per day, and more than 75% of that can happen standing up, using a specialized system of tendons and ligaments in the legs, sometimes called the stay apparatus, that locks the joints in place with minimal muscular effort. This lets a horse doze standing in three- to four-minute bursts throughout the day and night without expending energy to stay upright.

But standing sleep in horses is limited to light, non-REM stages. REM sleep requires the muscle atonia that defines the state, and a horse whose muscles went fully limp while standing would simply collapse. Horses need at least 30 minutes of REM sleep in each 24-hour period to stay physically and neurologically healthy, and that REM sleep can only happen lying down, usually in short bouts of a few minutes at a time overnight when the horse feels safe enough in its environment to become fully recumbent. Horses deprived of safe lying-down conditions, in an unfamiliar stable, a herd without established trust, or a paddock that feels exposed, can develop a measurable REM deficit, sometimes visible as buckling knees or brief collapse from micro-sleep episodes during the day, a condition equine veterinarians recognize and treat by addressing the horse's sense of security in its resting environment.

A horse can meet most of its daily rest needs standing up thanks to a locking leg-tendon system, but it still must lie down for at least 30 minutes of REM sleep every day or risk a measurable sleep deficit.

Newborn orcas: a month without sleep

Perhaps the most extreme documented case in mammalian sleep science comes from killer whales. In research led by UCLA neuroscientist Jerome Siegel and postdoctoral researcher Oleg Lyamin, published through observational studies of captive and wild orca populations, newborn orca calves and their mothers were found to go essentially without normal sleep for the first month after birth. Neither mother nor calf closed both eyes or stopped moving for more than a few minutes at a stretch during that period. Adult orcas outside this window typically rest 5 to 8 hours a night, floating motionlessly with both eyes closed in a pattern resembling unihemispheric sleep similar to dolphins.

The proposed explanation is a combination of survival pressures unique to newborn cetaceans. A newborn calf has almost no insulating blubber layer yet, so continuous swimming helps generate body heat and prevents dangerous heat loss in open water. Constant movement also keeps the calf near the surface to breathe reliably, since a newborn's breathing reflex is not yet fully automatic, and it reduces the calf's vulnerability to opportunistic predators like sharks during the most defenseless period of its life. What makes this finding especially striking to sleep researchers is the absence of any rebound. Once the calf's sleep pattern gradually normalizes over the following weeks, neither mother nor calf shows the kind of compensatory extra-deep sleep that sleep-deprived mammals typically display when finally allowed to rest. That absence of rebound challenges a core assumption in sleep science, that lost sleep must eventually be repaid, and remains one of the more genuinely puzzling findings in the field.

Insect and invertebrate sleep: pushing sleep's origin back further than expected

For a long time it was unclear whether invertebrates slept at all, which raised the possibility that sleep was purely a vertebrate invention tied to a complex, centralized brain. That idea has eroded steadily as researchers applied the same behavioral criteria, sustained inactivity, raised arousal threshold, and homeostatic rebound, to progressively simpler nervous systems.

Fruit flies (Drosophila melanogaster) are now one of the most important model organisms in sleep genetics precisely because they meet every criterion for true sleep. They show sustained inactivity that can be interrupted but requires "catch-up" sleep if prevented, a measurable rebound after deprivation, and a higher threshold for waking during their inactive period than during their active period. Fly sleep is regulated by many of the same signaling molecules involved in mammalian sleep, including adenosine, the chemical that gradually builds up in the brain during wakefulness and produces the sensation of sleepiness in humans, and that caffeine works by blocking. In some experiments, sleep-deprived fruit flies die faster than flies deprived of food entirely, underscoring how costly sleep loss is even in a nervous system with roughly 100,000 neurons, compared to the roughly 86 billion in a human brain.

Honeybees offer a different, more social angle on invertebrate sleep. A landmark 2010 study published in the Proceedings of the National Academy of Sciences (PNAS) by Barrett Klein and colleagues found that sleep-deprived worker bees produced measurably less precise waggle dances, the figure-eight movement pattern bees use to communicate the direction and distance of food sources to nestmates. Follower bees paying attention to a sleep-deprived dancer's imprecise signal were less accurate at locating the advertised food source afterward. That means sleep loss in a single bee can degrade the foraging efficiency of the entire colony, a rare documented case of individual sleep deprivation producing a measurable social, not just personal, cost in any animal.

Even more striking, a species of jellyfish-like organism called Cassiopea, the upside-down jellyfish, has been shown in controlled experiments to display sleep-like states: reduced pulsing activity at night, a higher threshold for a stimulus to trigger a response, and a homeostatic rebound in activity after being experimentally kept active overnight. Cassiopea has no brain at all, only a diffuse nerve net with no centralized control structure. That a sleep-like state exists in an animal without any centralized nervous system has led some researchers to propose that sleep, or something like it, may predate centralized brains entirely, existing as a basic maintenance function of nerve cells that was simply retained and elaborated as nervous systems grew more complex over evolutionary time.

Do insects and other invertebrates dream?

There is no confirmed evidence that insects experience dreaming in any sense comparable to mammalian REM dreaming. What has been established is that fruit flies and some other invertebrates show measurable sleep stages with different depths, and that sleep-deprived insects display cognitive and behavioral impairments, but the subjective experience, if any, during invertebrate sleep remains scientifically untestable with current tools.

REM sleep beyond mammals: birds, and the surprising case of the octopus

Rapid eye movement (REM) sleep, marked by muscle atonia, rapid eye movements, and, in humans, vivid dreaming, was long assumed to be a mammalian specialty. It has since been confirmed in birds, which show REM bouts, though typically shorter than mammalian REM, often lasting only a few seconds at a time. The real surprise came from a 2021 study published in iScience by Sylvia Medeiros, Sidarta Ribeiro, and colleagues, which documented a REM-like sleep state in the common octopus (Octopus vulgaris and later Octopus insularis).

Octopuses evolved along a completely separate branch of the animal family tree from vertebrates, with a last common ancestor with humans living more than 500 million years ago, long before centralized camera-like eyes or complex brains as we know them existed in either lineage. During the octopus REM-like state, the animal's skin, which it normally uses for active camouflage by rapidly changing color and texture, cycles through vivid color and pattern changes even though the octopus's eyes remain closed and it is clearly at rest. Researchers observed these color-change sequences resembling the octopus's waking camouflage displays and even prey-capture postures, suggesting the animal may be replaying recent experiences during this sleep state, a functional parallel to how mammalian REM sleep is believed to support memory consolidation through neural replay.

Because octopuses and vertebrates diverged so far back in evolutionary history, and because their brains are organized completely differently, with most of an octopus's roughly 500 million neurons distributed through its arms rather than centralized, the discovery of REM-like sleep in both lineages raises a genuine open question. Either REM sleep, or some functional precursor to it, is old enough to predate the split between vertebrate and invertebrate lineages entirely, or, more likely according to most researchers in the field, REM-like sleep evolved independently at least twice, a case of convergent evolution converging on a strikingly similar solution to whatever problem REM sleep solves.

What sleep deprivation reveals about why animals sleep at all

If sleep were a minor luxury, evolution should have eliminated it from at least some fast-moving prey species, since an unconscious or half-conscious animal is a vulnerable one. Instead, sleep, in some form, appears to be universal across every animal lineage studied carefully enough to test for it, from jellyfish-relatives to elephants. That universality is one of the strongest pieces of indirect evidence that sleep performs some function so fundamental to nervous system operation that no animal has evolved a way around it entirely, only ways to minimize, fragment, or redistribute it.

Direct evidence comes from sleep deprivation experiments, which are ethically fraught and largely restricted to laboratory animals, but which paint a consistent and severe picture. In classic experiments conducted by sleep researcher Allan Rechtschaffen at the University of Chicago in the 1980s and 1990s, rats subjected to total sleep deprivation using a rotating-disk apparatus died within roughly 2 to 3 weeks, a shorter survival time in some trials than rats deprived of food entirely. The cause of death was not a single organ failure but a systemic biological collapse: immune system breakdown, uncontrolled weight loss despite dramatically increased food intake, skin lesions, and body temperature dysregulation. That combination of symptoms suggests sleep is load-bearing for an unusually wide range of physiological maintenance functions simultaneously, not narrowly tied to one organ system or one biological process.

In classic sleep-deprivation experiments at the University of Chicago, rats denied all sleep died within 2 to 3 weeks, faster in some trials than rats denied food entirely, despite eating far more than usual before death.

Common misconceptions about how animals sleep

Myth: dolphins never sleep. This idea persists because dolphins never lose full consciousness the way a sleeping human does. In reality, dolphins sleep constantly through unihemispheric slow-wave sleep, resting one brain hemisphere at a time. The absence of total unconsciousness is not the absence of sleep; it is sleep organized differently across the two halves of the brain.

Myth: animals that sleep very little, like elephants and giraffes, need less sleep than other mammals. The available evidence suggests the opposite may be closer to true: wild elephants and giraffes appear to sleep far less than what would likely be optimal for full restorative benefit, and the gap between their wild and captive sleep durations, several-fold in both species, suggests environmental risk is suppressing sleep below a biologically preferred level rather than reflecting a genuinely lower requirement.

Myth: only mammals and birds have real, measurable sleep. Fruit flies, honeybees, and even the brainless jellyfish relative Cassiopea all meet the standard behavioral criteria for sleep: sustained inactivity, raised arousal threshold, and homeostatic rebound after deprivation. Sleep-like states appear to be far more evolutionarily ancient and widespread than researchers assumed even twenty years ago.

Myth: standing sleep and lying-down sleep are functionally interchangeable. In horses, standing sleep only supports lighter, non-REM rest stages. REM sleep specifically requires the muscle atonia that would cause a standing animal to collapse, so horses, like most large ungulates capable of standing sleep, still must lie down for a portion of their daily rest to get REM sleep, and deprivation of safe lying-down time produces measurable health consequences.

Species sleep duration comparison

Typical daily sleep duration and conditions across selected species, based on published field and laboratory studies.

Species	Typical sleep duration	Setting	Notable feature
Brown bat (Myotis lucifugus)	19–20 hours	Wild/captive	Among the longest documented mammalian sleep durations
Koala	18–22 hours	Wild	Low-energy eucalyptus diet requires minimal waking activity
Domestic cat	12–16 hours	Domestic	Burst-hunting physiology followed by long rest
Human (adult)	7–9 hours	N/A	Single consolidated nightly period is unusual among mammals
Horse	5–7 hours (mostly standing)	Domestic/wild	At least 30 min REM sleep must occur lying down
Captive African elephant	4–6 hours	Captive	No predation pressure
Wild African elephant	~2 hours	Wild	Least documented sleep of any land mammal studied
Wild giraffe	0.5–2 hours (fragmented)	Wild	Sleeps in bouts as short as a few minutes
Great frigatebird (in flight)	~42 minutes/day	In-flight, migration	Sleeps unihemispherically while gliding
Newborn orca calf	~0 for first month	Wild/captive	No sleep and no rebound documented in early weeks

FAQ

Q: Which animal sleeps the most? A: Among well-studied species, the brown bat sleeps about 19 to 20 hours per day, and koalas sleep 18 to 22 hours, primarily because their eucalyptus-leaf diet provides very little usable energy and requires minimal waking activity to process.

Q: Do fish sleep? A: Yes. Fish enter sleep-like states, reduced activity, altered posture, raised arousal thresholds, and homeostatic rebound after deprivation, even though they lack eyelids and their resting state looks like slow hovering rather than stillness. Zebrafish show both slow-wave and REM-like sleep stages, which has made them a widely used model for genetic sleep research.

Q: Why do cats sleep so much? A: Domestic cats sleep 12 to 16 hours per day because their evolutionary lineage as small-prey ambush hunters relies on short, intense bursts of energy followed by long recovery periods. Extended sleep supports muscle repair, neural consolidation, and immune maintenance between hunting bouts.

Q: Is it true dolphins never sleep? A: No. This myth persists because dolphins never become fully unconscious the way land mammals do. They sleep continuously through unihemispheric slow-wave sleep, resting one brain hemisphere while the other stays alert to control breathing and monitor the environment.

Q: What happens to animals that are prevented from sleeping? A: The outcomes documented in laboratory studies are severe. In classic experiments by Allan Rechtschaffen at the University of Chicago, totally sleep-deprived rats died within 2 to 3 weeks from systemic biological breakdown, including immune failure, uncontrolled weight loss despite increased eating, and organ dysfunction.

Q: How do wild elephants function on only 2 hours of sleep? A: This remains an open question in sleep science. Researchers led by Paul Manger found wild African elephants sleep about 2 hours per night, far less than captive elephants, and propose that predation risk and constant foraging demands compress sleep to a bare minimum. Whether this reflects unusually efficient sleep or an unmet biological need is not yet resolved.

Q: Can animals sleep while moving, flying, or swimming? A: Some can, using unihemispheric sleep. Great frigatebirds have been recorded sleeping in brief bouts while gliding during multi-day transoceanic flights, and dolphins swim slowly at the surface while one brain hemisphere rests. No confirmed case exists of an animal navigating complex terrain or hunting while fully asleep.

Q: Do insects really sleep, or are they just resting? A: Fruit flies meet all the standard behavioral criteria for true sleep: sustained inactivity, a raised threshold for waking, and a homeostatic rebound after being kept awake. Fly sleep is even regulated by some of the same molecules involved in human sleep, including the chemical adenosine, which caffeine blocks.

Q: Does REM sleep exist in animals other than mammals and birds? A: Yes. A 2021 study documented a REM-like sleep state in octopuses, complete with skin color-pattern changes resembling waking behavior, even though octopuses diverged from the vertebrate lineage more than 500 million years ago and have a completely different brain structure.

Q: Why don't newborn orcas need to catch up on lost sleep later? A: Researchers do not fully know. Newborn orca calves and their mothers go without measurable sleep for roughly the first month after birth, but once normal sleep patterns resume, neither shows the compensatory extra-deep sleep typically seen after sleep deprivation in other mammals, which challenges standard assumptions about how sleep debt works.

Q: Is standing sleep as restorative as lying-down sleep? A: Not entirely. Horses can meet most of their light, non-REM rest needs standing up using a locking leg-tendon system, but REM sleep requires muscle atonia that only happens lying down, so horses still need at least 30 minutes of lying-down REM sleep daily to avoid a measurable sleep deficit.

The takeaway

Sleep is not a single behavior that animals either do or do not perform. It is a family of overlapping neural states, slow-wave sleep, REM sleep, unihemispheric sleep, and possibly still-undiscovered variants, that evolution has stretched, compressed, split, and redistributed to fit an enormous range of survival pressures. What stays constant across every species studied so far, from jellyfish relatives to elephants to octopuses, is that some form of sleep appears to be non-negotiable, and going without it carries real, sometimes fatal, biological costs. The diversity is not evidence that sleep is optional for some animals; it is evidence of how many different ways nervous systems have found to protect a function too important to lose.

---

Read next:

• How Sleep Affects the Brain: The Science of Why You Need It
• Sleep Science — Why Your Brain Needs Rest
• Circadian Rhythm — Your Body's Internal Clock