��è����Ƶ

I am napping in a chimpanzee nest, 12 metres above the ground in Uganda’s Toro-Semliki Wildlife Reserve. Sleep comes easier than you might expect. The nest is surprisingly well-engineered: springy, cupped and stable in a way that feels less like a pile of branches than a purpose-built piece of furniture.

Then I wake up. I am suddenly, acutely aware that I am not in bed, but balanced high in the canopy.

I had climbed into the nest to investigate one of life’s least appreciated evolutionary problems: sleep, a paradoxical state that is essential yet renders the sleeper extremely vulnerable. In particular, I wanted to find out why human sleep is so weird. We do it far less than a primate like us ought to, so how did we evolve to need so little? Our sleep patterns are strange, too, with an unusually high proportion of the rapid eye movement stage.

By delving into humanity’s deep evolutionary past, as well as studying other primates, I hoped to find out why humans not only began to sleep less, but also reconfigured the way we do it. In the process, I discovered the long-overlooked role of sleep in humanity’s success, and found surprising answers to the question of whether we should aim to “palaeo sleep”, the way ancient humans did.

Sleep is wonderful, but also a terrible idea. It is biologically indispensable, critical for cognition, immune function, memory, attention and health. But a sleeping animal isn’t able to defend itself or watch for danger. Sleep also steals time that could be devoted to productive activities, such as finding a mate or foraging.

Subscriber-only newsletter

Sign up to Our Human Story

Each month, Michael Marshall unearths the latest news and ideas about ancient humans, evolution, archaeology and more.

Sign up to newsletter

That tension runs through the animal kingdom, meaning that species don’t evolve to maximise sleep, but instead end up with a compromise. Across mammals, sleep duration varies drastically between species, from 2 hours a day for to 20 hours for the pocket mouse. That disparity is the result of determined by predation risk, metabolic needs, foraging time and the safety of the sleeping site. Predators, for example, sleep longer than prey animals.

Imagine asking a biologist, with no knowledge of human sleep habits, to predict how much a primate like us should be getting. Give them our body size, brain size, reproductive schedule, place on the primate family tree and typical slumber environment, and they would . This is about 35 per cent more than the 7 hours that, after we control for type of society (small off-grid vs large on-grid), seems to be close to the human average. In fact, humans sleep less than any of the 30 primate species for which comparable data has been assembled.

This isn’t a small deviation that can be explained away by invoking cultural traditions or modern overstimulation. As I outline in my new book, , it is a deep evolutionary outlier, and any serious explanation of human evolution has to reckon with that fact.

The strangeness doesn’t stop with how little we sleep. It extends to how our sleep is built.

In humans and many other mammals, sleep cycles through two main phases: rapid eye movement sleep (REM) and non-rapid eye movement sleep (NREM). It is during the NREM phase that our deepest sleep occurs, associated with physical restoration, strengthening of the immune system and the consolidation of memories, especially declarative memories (those to do with recalling facts and events).

REM, by contrast, is associated with vivid dreaming, temporary muscle paralysis and wake-like brain activity that results in a paradoxical state in which the brain is highly metabolically active while many skeletal muscles are switched off. This phase appears to play an important role in emotional processing and integrating some memories and skills into brain pathways.

In REM sleep, thermoregulation – the process by which the body maintains its core temperature – is largely turned off because the brain temporarily prioritises dream-related neural activity and muscle paralysis over body-temperature control. So, in this state, body temperature is more dependent on the surrounding environment and, in some ways, the sleeper is maximally compromised.

Yet humans, despite sleeping less overall, devote a remarkably high proportion of their shorter nights to REM.

The remarkable tale of how humans nearly didn’t conquer the world

Over tens of thousands of years, waves of Homo sapiens set out across Europe and Asia, only for their societies and cultures to mysteriously vanish. At last, ancient DNA is revealing why

This finding came from a 2018 study I carried out with my colleague , who studies evolutionary medicine at Duke University, North Carolina, which compared the sleep characteristics of 30 primates, including chimps, orangutans, macaques and lemurs. We found that humans have the highest percentage of REM sleep among the primates sampled. That doesn’t mean we spend dramatically more total time in REM than expected. In fact, human REM duration was only modestly above the predicted average.

The real shift lies in reduced NREM duration. Our model predicted 8.4 hours of NREM for humans, but the actual time typically spent in this phase is 5.4 hours. We are, in effect, a primate that has compressed the night largely by cutting back on NREM while preserving a relatively large share of REM.

That is a very odd situation. If evolutionary forces resulted in reduced sleep time for a particular species, you might expect the most costly components to be shaved down first. But humans seem to have done almost the reverse. We became short sleepers without surrendering the proportion of sleep most associated with dreaming and a brain that remains metabolically active through the night.

From one angle, this looks almost reckless. Why would natural selection produce an ape that sleeps less, while protecting such a costly state? From another angle, it looks like a clue. Perhaps what was under selection wasn’t less sleep for its own sake, but a new kind of sleep economy: shorter, more efficient and carefully reweighted towards the cognitive functions that mattered most in our lineage.

That possibility becomes more plausible when we consider the places where humans sleep compared with other primates. Where you sleep matters: whether you can conceal yourself, whether you can cling safely, whether your group offers protection and how much danger will result from being unconscious in your environment.

Great apes solved some of these problems by building nests. Those sleeping platforms probably improved sleep quality by increasing physical comfort and reducing the danger posed by predators. Some favoured trees often selected by chimpanzees, like Cynometra or African ironwood, also . Barbara Fruth and Gottfried Hohmann of the Max Planck Institute for Animal Behaviour in Radolfzell, Germany, have even made the case that nest-building resulted in a for great apes by enabling a better night’s rest, which could have underpinned the development of object-manipulation skills.

Perched in that chimpanzee nest above Toro-Semliki, I begin to sense the scale of the transition our lineage eventually made. At some point in our deep past, our ancestors left this world of arboreal mattresses and moved towards something more precarious: sleeping habitually on the ground. In doing so, we became the most sleepless of them all.

Humans are the only primates that regularly snooze on terra firma across age and sex classes (although some male gorillas and some chimpanzees do also opt for the ground, where the risk of attack by predators is low). This move from the trees seems to have become a defining feature by the time of our ancestor Homo erectus around 1.8 million years ago.

From the perspective of a chimpanzee nest, the ground looks like a terrible idea. On the savannah or woodland floor, sleepers become more accessible to predators and more exposed to hostile rivals.

Human sleep paradox

This is why the human sleep paradox – the fact that our lineage seems to have been under selection to get it done faster, more efficiently and under riskier conditions than our closest relatives – is so interesting. It prevents us from telling a lazy story: that humans are tired because modern life is busy, or that our sleep troubles began with electricity, or that the 7-hour average is merely a cultural aberration.

The comparative record suggests something much deeper: that we are a lineage that departed sharply from the primate norm long before alarm clocks, shift work or glowing screens. The question isn’t whether modernity has damaged sleep, but what kind of evolutionary pathway could have produced such an unusual sleeper in the first place.

The answer, I think, begins where the nest in the trees ends.

Chimpanzees solved some of the problems of sleep by perfecting life in the trees. Humans solved it by abandoning trees and then remaking their nights on the ground. In doing so, they entered a far riskier situation that, over time, they transformed through cooperation, shelter, fire and shared vigilance into something no other ape had ever built: a social sleep niche. Only then does the oddity of human sleep begin to make sense.

At some point in the evolution of early Homo, the place where bodies slept became a collective technology: a defended patch of ground, a firelit perimeter, a shared shelter, a zone of lowered risk created by many individuals at once. That shift matters because sleep is never just a trait of a brain in isolation. It also depends on the environment. And in early humans, that environment increasingly became something they built together.

Deep sleep seems to lead to more eureka moments

After a nap, people who entered the second stage of sleep were more likely to spot a solution to a problem than those who slept lightly or not at all

One way to think about this is through what I call the sleep exophenotype. Biologists sometimes refer to the extended phenotype, the idea that the effects of an organism’s genes don’t stop at the skin. A beaver’s dam is the classic example: although the dam isn’t part of the beaver’s body, it is still an evolved expression of beaver behaviour that reshapes the animal’s world in ways that feed back into survival and reproduction.

Human sleep sites belong in that same family of ideas, but with a specifically social twist. A sleep exophenotype is the external sleep environment that organisms create, modify or inherit in ways that alter the quality, safety, timing or expression of sleep. In humans, this exophenotype was communal. It was built out of bodies, technologies, habits and mutual dependence.

To describe this, I use the acronym SHELL, which is the suite of protective human sleep innovations: shelter, hearth, environmental preparation, light and lookouts. Humans wrapped sleep inside a socially engineered shell. This idea grew, in part, out of my work with the Hadza, a hunter-gatherer society in Tanzania whose way of life can offer clues about the kinds of sleep environments humans may have experienced before agriculture and industrialisation (see “Should we “palaeo sleep”?“) .

My research assistant Ibrahim Mabulla and I lived and collaborated with the Hadza for a summer and used wearable devices to measure their sleep. We learned that group sleeping reduced the odds that every individual would be equally vulnerable at once, with people waking at different times across the night, helping to distribute vigilance over that time. In fact, there were only 18 minutes across 20 nights in which all the group members were asleep simultaneously. At least one person was awake at almost any given time throughout the night.

Human sleep, in other words, appears to have become distributed across the group. Shelters and hearths provided warmth, deterrence and a social centre of gravity. Environmental preparation – clearing ground, organising bodies, reusing favoured camp locations and collectively shaping the wider sleep site – made nighttime spaces more predictable and defensible. Light from fire altered both visibility and evening sociality, extending the human niche into the dark, and grass huts buffered nighttime thermal variation. In short, we changed the ecology around sleep.

It’s your perception of sleep that’s making you feel tired all day

How we feel about a night’s sleep can have a bigger impact on mood and grogginess than actual hours of rest. Here’s how to change your mindset to feel more energised

Once sleep became embedded in that shell, social sleep could do something extraordinary: it could make terrestrial sleep workable, accommodating ever-increasing group sizes over evolutionary time.

This may be one of the most underappreciated transitions in human evolution. We often celebrate bipedalism, fire and language as the great enabling revolutions of our lineage. But social sleep belongs on that list. Yes, some other social mammals, such as baboons, also benefit from group sleeping and collective vigilance, but the human case is unusually elaborate.

A protected sleep shell would have allowed early humans to gain the , such as improved creativity and innovation, without paying the full costs of vulnerability.

Once early humans could reliably manufacture warmer, quieter and safer nocturnal microhabitats, they were no longer bound to locations close to the equator in quite the same way. All this would have been a crucial pre-adaptation to prepare humans for migrating to latitudes away from equatorial Africa and eventually thriving in almost every ecology on Earth. Before humans colonised new continents, we first had to learn how to carry a habitable night with us.

Seen this way, the sleep site isn’t a footnote to human evolution. It is one of its hidden engines, creating a nighttime world within which humans could survive and then prosper. Out of that shell came not just safer sleep, but also a platform for cognition, innovation and expansion. We became modern humans not only by changing our bodies and brains, but by constructing external niches that let those bodies and brains do something unprecedented: sleep briefly, sleep deeply, sleep socially and wake ready to build the incredible world we now inhabit.

��