How did dinosaurs sleep?

Dinosaurs slept, but a fossil almost never preserves sleep itself. It can record an articulated skeleton folded into a compact position, the imprint of an animal that sat down, or an adult settled over a clutch. Such evidence reconstructs posture. Eye closure, brain state, sleep duration and dreams do not fossilise.

The strongest examples are small predatory dinosaurs found with legs tucked beneath the body, the tail wrapped around it and the head resting near a forelimb. Two specimens of Mei long are particularly important. The same elaborate arrangement in separate individuals is difficult to explain as random collapse, current action or burial pressure. It resembles the resting posture of modern birds and shows that this way of folding the body preceded birds in their modern form.

What can a palaeontologist reasonably call sleep?

Sleep is a state of an active brain. In living animals it is recognised through responsiveness, eye movement, muscle tone and electrical activity in the nervous system. A fossil skeleton preserves none of those measurements. “Sleeping dinosaur” therefore names the most economical interpretation of a posture, not a direct physiological reading.

Confidence rises with anatomical detail. An articulated skeleton, with each limb still in place, is stronger evidence than scattered bone. Repetition of an unusual pose in several individuals is more persuasive than one specimen. Rapid burial matters because water, scavengers and decay can rearrange a carcass. Comparison with birds is useful only when corresponding joints and muscles can actually produce the same configuration.

This creates a hierarchy. Bone position is observed. Prolonged rest can be inferred. Sleep is likely when the posture is economical, repeated and inconsistent with transport. A particular sleep phase, the reason for waking or a subjective experience remains unreachable.

Mei long and the bird-like pose

The small troodontid Mei long was described from Early Cretaceous deposits of Liaoning in 2004. Its name means “sleeping dragon”. The skeleton lay over folded hind limbs. Its long tail curved around the body, the neck bent backwards and the head rested close to the forelimb. The entire animal occupied a compact patch despite having a neck and tail that could extend much farther.

One spectacular skeleton might be an accident. A second specimen, published in 2012, preserved almost the same arrangement. It again tucked its limbs, brought the tail forward beside the trunk and turned the head towards the shoulder. Repetition makes it much less likely that convulsion, current action or rock compression created the first pose from an unrelated carcass.

The general conclusion is more secure than every detail: small maniraptoran dinosaurs could fold their bodies much as birds do. Saying the head was “under a wing” is only shorthand. The feathered forelimb of a non-avian troodontid was not yet the same structure as a modern bird's wing.

Why curling up was useful

Tucked limbs do not project sideways and need not be held by continuously tense muscles. A tail wrapped around the trunk reduces occupied space, while turning the neck lets the head rest against the body or forelimb. This configuration suits prolonged rest and protects vulnerable parts.

Compactness also reduces exposed surface area and therefore heat loss. That could matter to a small animal with a high surface-to-volume ratio on a cold night. Feathers improved insulation, although their exact density cannot be recovered for every sleeping specimen. Our guide to feathers, skin and colour distinguishes direct impressions from coverage inferred through relationships.

Heat conservation is a functional explanation, not a weather report for the night an individual died. Air temperature, bedding and the depth of actual sleep remain unknown. Rest, protection and warmth could all favour the same pose without one exclusive cause. The wider metabolic context is discussed in dinosaur thermophysiology.

Other dinosaurs preserved in comparable positions

Mei long is the clearest example, not the only one. The Mongolian troodontid Sinornithoides youngi lay compactly with tucked hind limbs, folded forelimbs and its head close to a hand. Preservation is less expressive than the two Mei specimens, but the basic arrangement is similar.

In 2023 the Mongolian alvarezsaurid Jaculinykus yaruui was named from a nearly complete articulated skeleton. Its limbs were tucked, the neck curved and the tail wrapped around the body. Alvarezsaurids sit outside Troodontidae, so this specimen broadened the distribution of the bird-like resting posture among maniraptorans.

Hypnovenator matsubaraetoheorum, named in 2024 from Japan, also preserved articulated parts with bent hind limbs and forelimbs placed nearby. It comes from river deposits rather than precisely the same environment as the Chinese specimens. One special sedimentary process therefore cannot explain every known example.

All these skeletons belong to relatively small bipedal theropods. They document one posture well but cannot be projected onto every dinosaur. Ceratopsians, ankylosaurs and multi-tonne sauropods had different body geometry and likely used different comfortable positions.

A theropod left a trace of sitting down

A skeleton fixes an endpoint; a track surface can preserve a sequence. An Early Jurassic trace complex from Utah, about 198 million years old, records a theropod walking, lowering its pelvis, placing both hands and bringing the tail into contact with the ground. It then rose and continued.

The trace preserves anatomical details. Feet stood symmetrically to either side of the body, lower legs approached the surface and palms faced inward. This agrees with theropod wrist construction: the hands could not rotate flat down in the old “bunny hands” pose.

This surface proves sitting and a pause, not sleep. The animal might have rested, watched its surroundings, sheltered from wind or stopped for another reason. Behavioural tracks are valuable precisely because their most secure claim can be narrower than the scene we imagine.

Brooding a clutch is not the same as sleeping

Oviraptorids provide one of the strongest fossil records of parental behaviour. Adult skeletons occur above ring-shaped clutches, with hind limbs at the centre and forelimbs spread over the eggs. This geometry closely resembles bird brooding and is difficult to explain as a carcass randomly washed onto a completed nest.

These specimens show that an adult occupied the clutch for an extended period and covered it with body and forelimbs. Associated eggs and embryos strengthen the behavioural interpretation.

A brooding animal can be awake. Birds doze on nests but also turn eggs, monitor threats and change position. A fossil cannot separate those minutes. The skeleton demonstrates brooding and a posture suitable for prolonged rest; sleep remains a possible episode rather than the recorded behaviour.

Where dinosaurs may have rested

A death site is not always a chosen sleeping place. Water can transport a body and decay can alter its position. The compact, articulated Mei skeletons indicate rapid burial. A shallow shelter and sudden sediment associated with volcanic activity have been suggested, but the precise cause of death remains unresolved.

Some dinosaurs genuinely used burrows. Adults and juveniles of small ornithopods have been recovered inside fossil tunnels. A burrow buffers temperature and hides its occupants, making it a suitable rest site. Yet even a skeleton inside a tunnel does not necessarily preserve a sleeping individual.

Forest litter, a dry bank, the middle of a herd or an open plain rarely leaves a recognisable bed. The record of dinosaur rest is therefore biased towards places where sediment quickly covered a body or retained a contact trace. The shortage of examples among large species reflects preservation as well as behaviour.

Did large dinosaurs sleep standing or lying down?

No articulated Tyrannosaurus, Triceratops or sauropod skeleton is generally accepted as an animal killed during sleep. Joints and tracks constrain possible positions but do not choose a nightly routine. A large dinosaur might doze standing, lower its chest and limbs for deeper rest, or alternate between the two.

Living elephants and giraffes show that body size does not dictate one answer. They can rest upright yet lie down for part of sleep. That is a biomechanical comparison, not a ready-made sauropod model. Columnar limbs, a long neck, air sacs and a distinctive centre of mass formed a different system.

A common claim says a multi-tonne dinosaur could not rise after lying down. Skeletons do not support that universal prohibition. Large quadrupeds had flexing joints, but the exact lowering and rising sequence is unknown for most genera. A reconstruction must test joint ranges and loads for a particular body rather than infer inability from mass alone.

Did dinosaurs sleep at night?

Night is not the preferred rest period of every animal. Some species are active by day, others in darkness, and many alter their schedule with season, temperature and food. A fossil pose contains no clock. The dusk on this page's cover creates a readable scene but does not classify Mei as an exclusively nocturnal sleeper.

Researchers have estimated activity patterns from the scleral ring and orbit dimensions. Comparisons with living birds and lizards produced possible daytime, night-time and irregularly active dinosaurs. The method is useful but sensitive to sampling, reconstructed measurements and how closely extinct eyes match modern ones. Large eyes alone do not prove a nocturnal hunter.

Season could reshape the schedule. High-latitude dinosaurs experienced very long summer days and winter darkness. Migration, food availability and internal heat retention affected activity. Fossils reveal constraints, not an hourly sleep timetable.

What birds and crocodilians contribute

Birds are living dinosaurs, so their sleep provides more than a superficial resemblance. Slow-wave and rapid-eye-movement states can be measured, while head position, muscle tone and eye state change. Ostriches may hold the neck upright and appear alert during part of sleep. Posture alone can be ambiguous even when the animal is alive.

The connection is explored in the guide to early birds and flight. Shared ancestry makes corresponding folds of limbs and neck evolutionarily meaningful, but modern abilities cannot simply be assigned to the common ancestor. Phase duration, temperature control and hemispheric activity may have changed repeatedly.

Crocodilians represent the other living archosaur branch. Young crocodiles in experiments often kept one eye directed towards a possible threat while closing the other, behaviour consistent with unihemispheric sleep known in some birds. Without a complete set of brain measurements the interpretation remains cautious. Even an ancestral capacity would not prove the same pattern in every dinosaur.

Dreams, snoring and hours of sleep are beyond the fossils

Rapid-eye-movement sleep in birds and mammals is associated with distinctive brain states, but a dream cannot be verified by simply watching a sleeping animal. Dinosaurs leave no electrical activity, eye motion or waking response. The claim that they dreamed is biologically plausible but not a palaeontological result.

Snoring depends on soft respiratory tissues and head position. Those structures seldom fossilise and sound leaves no trace. Mass, metabolism and lifestyle can suggest comparisons, yet they cannot calculate how many hours Tyrannosaurus slept or how often a sauropod woke.

Observed evidence, inference and art

Bone position, articulation, body-contact traces, an adult's placement above eggs and sedimentary setting are direct observations. Repeated specimens reveal recurring postures. Rest, rapid burial and the function of compactness are inferences of different strength. Sleep is a strong explanation for Mei long, not a measured state.

Night-time, warmth beneath feathers and the selected shelter require separate arguments. Colour, bedding, facial expression and exact light belong to artistic reconstruction. The cover combines an evidence-based maniraptoran posture with an imagined place and moment.

Conclusion

The best evidence for dinosaur sleep is not a closed eye but a repeated compact posture among small theropods. Two Mei long, Sinornithoides, Jaculinykus and Hypnovenator show that tucked limbs, a tail around the body and a head turned towards the forelimb formed a real mode of prolonged rest. The Utah track confirms that a theropod could lower its whole body, although the pause itself does not demonstrate sleep.

Nest skeletons document brooding, burrows document shelter and eye anatomy constrains likely activity periods. None reveals sleep duration, phases or dreams. Dinosaur resting positions can therefore be reconstructed with surprising detail while the internal state of the animal remains inaccessible.