Why is the night sky dark?

Foundations

The obvious answer

Ask why the night sky is dark and you will get the obvious answer: space is mostly empty, and the stars are far away. Obvious, and wrong. Johannes Kepler saw the problem in 1610, Edmond Halley worried at it in 1720, and a doctor in Bremen named Heinrich Olbers gave the puzzle its name in 1823. It is one of the deepest questions a child can ask, and the true answer was not understood until the twentieth century.

Why distance does not help

Picture the stars arranged in shells around you, like the layers of an onion. A star in a shell twice as far away looks a quarter as bright. But a shell twice as far away has four times the area and, in a universe that goes on for ever, four times as many stars. The two effects cancel exactly. Every shell adds the same amount of light to your sky.

Add up an endless number of shells and the total has no limit. In practice something almost as strange would happen instead. In a forest that is large enough, every line of sight eventually ends on a tree trunk, however thinly the trees are spread. In an endless, unchanging universe of stars, every line of sight eventually ends on the surface of a star, and the whole sky, day and night, would blaze like the surface of the Sun. It does not. So something in “endless” or “unchanging” must be wrong.

The fixes that failed

Perhaps dust between the stars soaks up the light? No. Dust that absorbs starlight heats up, and given enough time it glows just as brightly as what it absorbs. Perhaps the universe is finite, with an edge? Kepler thought so, for exactly this reason, but an edge raises more questions than it answers. Perhaps the stars are arranged in clusters within clusters, with emptiness between? That can be made to work on paper, and it is not how the universe looks.

The poet’s answer

In 1848 Edgar Allan Poe, in a strange prose poem called Eureka, wrote down the answer that would eventually hold. If the succession of stars were endless, he reasoned, the sky would be uniformly bright, so the only way to understand the darkness is to suppose that the distant background is so far away that no light from it has yet been able to reach us. The universe, in other words, has not been here for ever. Light takes time to travel, and there has not been enough time.

Lord Kelvin made the same point with numbers in 1901. Stars do not shine for ever either, and a universe of stars with finite lifetimes cannot fill space with light.

The real answer, in numbers

For a line of sight to be blocked by a star it must travel, on average, something like a hundred thousand billion billion light-years, because stars are small and space is very empty. Light has been travelling for at most 13.8 billion years. So only about one line of sight in ten trillion has yet reached a star. The rest end in darkness, or rather, they end in the early universe, before any star existed.

The sky is dark because the universe is young. Stars have shone for only about 13 billion years, and no star shines for much longer than that.

What about the expansion?

The universe is also expanding, and expansion stretches light from distant galaxies to longer wavelengths and lower energies, dimming it. The effect is real, and it makes the night sky darker than it would otherwise be by about forty percent. But it is a supporting actor. A universe of the same age that was not expanding at all would still have a dark night sky. Age, not expansion, is the reason.

The twist: the sky is bright after all

Here is the part Olbers could not have guessed. Every line of sight does end on something glowing. Follow any direction far enough back and it reaches the moment, 380,000 years after the Big Bang, when the universe was a fog at three thousand kelvin, as bright as the surface of a cool star. That wall of light surrounds us; it is the afterglow of the next lesson. We cannot see it with our eyes because the expansion has stretched it eleven hundred times, from the orange glow of a flame into microwaves at 2.7 degrees above absolute zero.

Olbers’s bright sky exists. It is just not made of stars, it is far older than any star, and it has been redshifted into invisibility. Which is a second answer, from the same dark sky: the universe is both finite in age and expanding.

Look up

The darkness between the stars is not emptiness. It is a measurement. It tells you that the universe had a beginning, not so very long ago, and has been growing ever since. Everything else in this course follows from that.

Deeper

The shell argument as an integral

Suppose stars of radius R and luminosity L fill space uniformly with number density n. A shell at distance r and thickness dr contains 4πr²n dr stars, each delivering a flux L/4πr², so the shell contributes nL dr, independent of r. The total grows without limit as the distance does. It is capped only when stars begin to hide one another: the mean distance a line of sight travels before meeting a stellar disc is 1/(nπR²), the background limit. For the average density of stars in the universe, roughly a hundred million Sun-like stars per cubic megaparsec, this works out to be of the order of 10²³ light-years. Beyond that, the sky would be fully tiled with stellar surfaces and would have the surface brightness of a star: at 5,800 kelvin, about twenty million watts per square metre per steradian.

Finite age

The universe has been transparent for 13.8 billion years and has had stars for about 13.6 billion. Only the sky within that lookback distance contributes, a fraction of roughly 10⁻¹³ of the background limit. The measured extragalactic background light, the summed glow of all galaxies, comes to about 50 nanowatts per square metre per steradian across the optical and infrared, half of it direct starlight and half starlight absorbed by dust and re-emitted. That is several hundred trillion times fainter than the Sun’s surface, in fair agreement with the crude estimate. The sky is dark to about the degree the age of the universe says it should be.

Kelvin’s energy argument

Even an eternal, static universe could not do it. Filling space with radiation at a stellar surface temperature needs an energy density of about 0.9 joules per cubic metre at 5,800 kelvin. Converting all the ordinary matter in the universe to energy at the efficiency of nuclear fusion, about 0.7 percent of its mass, yields about 3 × 10⁻¹³ joules per cubic metre, a trillion times too little. Stars do not have the fuel. The sky is dark because the universe is young and because stars are finite, and both facts are needed.

Expansion: surface-brightness dimming

In an expanding universe the surface brightness of a source at redshift z falls by (1+z)⁴: one factor for the energy lost by each photon, one for their reduced arrival rate, and two for the apparent enlargement of the source. Integrated over realistic galaxy populations, Wesson and collaborators found that expansion reduces the background to about 0.6 of what a static universe of the same age would give. Real, but secondary. The finite lifetime of the galaxies sets the order of magnitude.

What you actually see from Earth

From the ground, the darkness is not even cosmic. At the best dark-sky sites the faint glow of the night sky is dominated by airglow, light emitted by excited molecules in the upper atmosphere, and by zodiacal light, sunlight scattered by dust within the solar system. Together they outshine the integrated light of all the stars and galaxies. To measure the cosmic optical background directly you have to leave. The New Horizons spacecraft, beyond Pluto where the zodiacal light is negligible, did exactly that.

The bright sky, redshifted

The afterglow completes the picture. Every line of sight ends on the surface of last scattering at a redshift of about 1,090, where the universe glowed at 3,000 kelvin. Dimming by (1+z)⁴, a factor of more than a trillion, and the shift into microwaves make it invisible to the eye, but it is there in every direction, and it carries about twenty times more energy per cubic metre than all the starlight ever emitted. The sky is, in the end, uniformly bright. Olbers was asking the right question about the wrong light.

Frontier

How dark, exactly

In 2022 the New Horizons team reported a direct measurement of the cosmic optical background of 16.4 ± 1.5 nanowatts per square metre per steradian, about twice what the known galaxies can supply, and an unexplained excess of about 8. A 2024 re-analysis with sixteen fields and a better model of the Milky Way’s own scattered light lowered the total to 11.2 ± 1.7, of which 8.2 ± 1.2 comes from catalogued galaxies, leaving 3.0 ± 2.0 formally unexplained. The anomaly has shrunk but not vanished. Candidates for what is left include galaxies too faint to count, stars flung into the space between galaxies, and, more speculatively, light from decaying dark matter. The James Webb Space Telescope, by counting the earliest and faintest galaxies, is closing the gap from the other side. This is active research.

The vanishing evidence

The darkness that carries all this information is itself disappearing. The 2016 world atlas of artificial sky brightness found that more than 80 percent of the world’s population lives under light-polluted skies, and that the Milky Way is hidden from more than a third of humanity. In the fast-growing cities of India the sky that Kepler and Olbers puzzled over is becoming something a student has to travel to see. A course about the whole universe should say so: the first piece of evidence in it is worth protecting.

Transcript and shot list

Target runtime about 7 minutes 30 seconds at a measured narration pace, with room for the visual beats to breathe. Shot types: ANIM is rendered from code with Manim, AI is a generated shot from Higgsfield with the shared style reference, REAL is agency imagery with credit, CAM is the narrator. Timecodes are targets and will move to match the recorded narration.

[00:00] S01 · AI · A child on a rooftop under a dark, star-filled sky, looking up; a town’s lights low on the horizon.

Ask why the night sky is dark and you’ll get the obvious answer. Space is mostly empty. The stars are far away. It’s obvious, and it’s wrong, and the first person to see that it was wrong was Johannes Kepler, in 1610. The right answer took another three hundred years, and it turns out to be one of the biggest facts about the universe, sitting in plain sight every night.

[00:35] S02 · ANIM · Cosmic calendar; the marker sits at midnight on 31 December; a zoom into the last second shows Kepler’s 1610 less than a second before the end.

On the cosmic calendar, Kepler asked the question less than a second ago. The answer is as old as the calendar itself.

[00:50] S03 · ANIM · Shells of stars around an eye; a shell twice as far has stars a quarter as bright and four times as many; the light from each shell tallies to the same bar.

Here’s why the obvious answer fails. Picture the stars arranged in shells around you, like the layers of an onion. A star in a shell twice as far away looks a quarter as bright. But a shell twice as far has four times the area, and four times as many stars. The two effects cancel, exactly. Every shell adds the same amount of light to your sky. And if the universe went on for ever, there would be no end of shells.

[01:35] S04 · ANIM · A thin forest seen from inside; the camera turns; every line of sight ends on a trunk; the trunks become stars and the whole frame turns Sun-bright.

Think of standing in a forest. If the forest is big enough, every direction you look ends on a tree trunk, no matter how thinly the trees are spread. In an endless, unchanging universe of stars, every direction you look would end on the surface of a star. The whole sky, day and night, would blaze like the surface of the Sun. It doesn’t. So something in “endless”, or something in “unchanging”, is wrong.

[02:20] S05 · ANIM · Three tried fixes, each crossed out: a dust cloud that heats until it glows; a universe with an edge; a hierarchy of clusters.

People tried fixes. Dust soaking up the light? No: dust that absorbs starlight heats up and eventually glows just as brightly. A universe with an edge? Kepler thought so, and an edge raises more questions than it answers. Stars in clusters within clusters, with emptiness between? It works on paper, and it isn’t how the universe looks.

[02:55] S06 · ANIM · Typographic: 1848, Poe, “no ray from it has yet been able to reach us”; then 1901, Kelvin, a star with a burning fuse.

The answer came from a poet. In 1848 Edgar Allan Poe wrote that if the stars went on for ever the sky would be bright, so the only way to understand the darkness is to suppose that the distant background is so far away that no light from it has yet been able to reach us. The universe has not been here for ever. There hasn’t been time. In 1901 Lord Kelvin put numbers on it, and added the other half: stars don’t shine for ever either.

[03:40] S07 · ANIM · A line of sight extends from the eye past sparse stars; a counter runs to ten to the twenty-three light-years before it hits one; a second line stops at 13.8 billion and ends in a red-orange glow.

Here are the numbers. Stars are small and space is very empty, so for a line of sight to hit a star it has to travel, on average, something like a hundred thousand billion billion light-years. Light has been travelling for thirteen point eight billion years. So only about one line of sight in ten trillion has reached a star. The others end in darkness. Or rather, they end in the early universe, before any star existed. The sky is dark because the universe is young.

[04:30] S08 · ANIM · Two identical skies side by side, labelled static and expanding; the expanding one dims to sixty percent.

What about the expansion? It’s real: the stretching of space dims light from distant galaxies, and it makes the sky about forty percent darker than it would otherwise be. But it’s a supporting actor. A universe this age that wasn’t expanding at all would still have a dark night sky. Age does the work.

[05:00] S09 · REAL · A New Horizons image of a dark field beyond Pluto (NASA, JHUAPL, SwRI; confirm credit). S10 · ANIM · A brightness scale from the Sun’s surface down fourteen powers of ten to the cosmic background; airglow and zodiacal light sit above it.

How dark is dark? Add up every star and galaxy in every direction and you get a glow several hundred trillion times fainter than the surface of the Sun. It’s so faint that from the ground you can’t even see it: the glow of Earth’s own upper atmosphere, and sunlight scattered off dust in the solar system, both outshine it. To measure it, the New Horizons spacecraft had to go beyond Pluto and look out from there. Its number, and a small excess in it that nobody has fully explained, is active research.

[05:50] S11 · ANIM · Every line of sight from the eye extends outward and stops on a glowing spherical wall; the wall’s colour slides from orange-white to invisible microwaves; the readout shows 3,000 K then 2.7 K.

Now the twist Olbers couldn’t have guessed. Every line of sight does end on something glowing. Look far enough in any direction and you reach the moment, three hundred and eighty thousand years after the Big Bang, when the whole universe was a fog at three thousand degrees, as bright as the surface of a cool star. That wall of light surrounds us. We can’t see it because the expansion has stretched it eleven hundred times, from the glow of a flame into microwaves at two point seven degrees above absolute zero. Olbers’s bright sky exists. It just isn’t made of stars, it’s older than any star, and it’s been redshifted out of sight. Same dark sky, second answer: the universe has a beginning, and it’s expanding.

[06:45] S12 · ANIM · Three labels: ESTABLISHED for age as the cause, ACTIVE RESEARCH for the background-light excess; then a world map of light pollution with a third of humanity shaded.

The scorecard. That the sky is dark mainly because the universe is young: established, and it has been since Kelvin. Exactly how dark, and whether a little unexplained light remains: active research. And one more thing worth knowing: more than a third of humanity can no longer see the Milky Way at all, because of our own lights. The first piece of evidence in this whole course is disappearing.

[07:15] S13 · AI into ANIM · Back to the rooftop; the town’s lights dim; the Milky Way comes out; the calendar returns.

So tonight, if you can, find somewhere dark and look up. The darkness between the stars isn’t emptiness. It’s a measurement. It says the universe had a beginning, not so very long ago, and has been growing ever since. Everything else in this story follows from that.

Shot list

Two AI shots, ten code-rendered shots, one agency image. Estimated Higgsfield use: 3 to 6 generations including retries.