The cosmic web is insultingly big. To state the bare fact that it's the single largest pattern found in nature does supreme disservice to the word largest. If there were a superlative to express a quantity greater than the greatest, that adjective might come close to addressing the magnitude of the cosmic web.

Think of the largest thing you possibly can. A planet? The Earth is so large that even though it's round, it appears flat in your backyard. A solar system? NASA's New Horizons probe traveled at thirty-six thousand miles per hour and took nine and half years just to make the hop to Pluto. A galaxy? A bustling stellar megalopolis, home to hundreds of billions of stars and a hundred billion suns’ worth of gas.

The cosmic web is made of galaxies, the same way that your body is made of cells. But even that metaphor breaks down—as do all metaphors—when describing the cosmic web. The cosmic web is made of galaxies, the same way your body is made of cells…if your cells were a million times smaller than they are.

But the galaxies themselves are only representatives of the true bones of the web: dark matter. The matter in our universe appears (or doesn't—ha, sorry, cosmology joke) to be made of some nonluminous particle or family of particles, one that doesn't interact with light or really anything. It's the dark matter that began pooling together billions of years ago; the “normal” matter simply fell into the already-existing gravitational valleys. When we see a galaxy, we must imagine a “halo” of dark matter surrounding it; likewise for a gigantic cluster. When we see the web, the light-emitting galaxies are tracers of the true structure underneath. Thus while everything I'll talk about below concerns stars and galaxies, keep in mind that those are only the metaphorical tips of the cosmic icebergs.

Let's start with some baby steps and work our way up. Voyager 1, launched in the late 1970s on a grand tour of the outer planets of the solar system, finally penetrated the bubble of our sun's influence, as defined by the boundary where the stream of charged particles racing outward from the sun's surface begins to mix with the general galactic milieu, in 2012. In three hundred years, that little spacecraft—which is no bigger than a small car, mind you—will reach the inner boundary of the Oort cloud, a thin, diffuse shell of frozen debris left over from the formation of the solar system.

Voyager 1 now enjoys the privilege of being the only humanmade object in interstellar space, the long gulfs of emptiness between the stars that make up the Milky Way. It will eventually pass by another star, coming within 1.6 light-years of Gliese 445, an unremarkable red dwarf currently situated about 18 light-years from the sun.

In forty thousand years.

While it's difficult to predict exactly, astronomers are pretty sure that's the closest Voyager 1 will come to another star. Ever.

In 230 million years, traveling at a steady thirty-eight thousand miles per hour, it will complete a single circumnavigation of the Milky Way without meeting anything larger than a stray bit of dust.

Here's another perspective. The sun's nearest neighbor is Proxima Centauri, another unremarkable red dwarf (they're rather common) about four light-years from our home. If you were to build a scale model of our galactic neighborhood, and you were to put the Earth a scant three feet away from the sun, Proxima Centauri would be two hundred miles away.

And we're just getting warmed up.

The Milky Way galaxy itself is around one hundred thousand light-years across. Simple math reveals that you could fit twenty-five thousand sun-Proxima distances across its breadth. In our scale model, with the sun three feet from the Earth and Proxima two hundred miles from that, the Milky Way in its entirety would stretch five million miles, which would put the edge about twenty times farther than the moon.

That's a big model.

Let's revisit our exploration of the deep universe using the handy new phrasing we learned when we first encountered Hubble and his fantastic result: the parsec, the quintessential astronomical jargon word. Like this: the Milky Way is about thirty thousand parsecs, or thirty kiloparsecs, across. It's still unimaginably gigantic, but at least I don't have to type a bunch of “illions” anymore.

The Andromeda Galaxy, the nearest major neighbor to our own galaxy, is about a megaparsec, or million parsecs, away from us. That simple statement hides an intriguing fact. Although galaxies are tens of thousands of times bigger than their constituent solar systems, the distances between galaxies are only a few times larger than galaxies themselves, making them sort of close together. Relatively speaking.

But the most interesting thing about the large-scale structure of the universe on scales bigger than even galaxies, and the reason it has a name like cosmic web, is that galaxies aren't just arranged randomly throughout the cosmos, a fact that quite surprised the early cosmographers—the mappers of the universe.

But in order to reveal the structure of the universe, you have to go much larger than our local patch. The work of Hubble and others to establish the expansion of the universe relied on a comparatively nearby sample of galaxies—not nearly enough to reveal anything other than a scattered smattering of so-called spiral nebulae. Sure, there were dense beehives of galaxies like the nearby Coma Cluster, which flirtatiously suggested the existence of greater structures, but for decades cosmologists weren't sure if that was a significant feature or just a random collection.

But beginning in the 1970s, astronomers began to perfect the technology necessary to systematically survey the locations of galaxies far from the familiar, using a potent combination of improved telescopes and computerized search algorithms. It was like the classic duo of scope + camera that transformed our understanding (or lack thereof) of the nineteenth-century universe, but on steroids. The new surveys pushed both wider and deeper, creating the first-ever maps of our universe on a truly universal scale. Galaxy by galaxy, insignificant point of light by insignificant point of light, each dot representing a hundred billion warm nuclear hearths set against the deep, vast coldness of our cosmos, structures began to appear.1

Long, thin ropes. Broad walls. Dense knots. Immense voids.

It was a web. A web made of galaxies. One by one, survey target by survey target, astronomers realized that at the largest scales, the galaxies of our universe are not scattered like salt sprinkled on a table. There was a pattern. There was order. The cosmic web is, and I'll repeat myself in case you missed it earlier, the single largest pattern found in nature. It stretches from one end of the observable universe to the other—and most likely beyond that, too, though we'll never see it.

The complexity of the web is astounding. The same substructures and intricacies you would find in a spiderweb are found at these vast scales. The filaments are long and tenuous, in some case only a few galaxies wide but stretching for dozens of megaparsecs. Some walls are almost impossibly large. The first one was simply called the Great Wall, only a few megaparsecs thick but hundreds of megaparsecs across.2 It was soon given a more specific name (the CfA2 Great Wall, for the curious, named after the survey where it was discovered) because it turns out there are a bunch of Great Walls in our universe.

The walls and filaments intersect at the clusters of galaxies, the largest gravitationally bound structures in the universe. Home to a thousand galaxies or more, they are relatively compact—a bare megaparsec or two from edge to edge. Most clusters are isolated creatures, settling into hydrostatic equilibrium and spherical symmetry long ago, but a few are cosmic car accidents, giving clues to their formation and constituents (there's a gruesome analogy here that I won't pursue further), like the infamous Bullet Cluster. Inside the clusters we find not just galaxies but also a hot, thin gas: the tenuous plasma of hydrogen, weak enough to register as a vacuum in the laboratory but hot enough to emit X-rays.

And between all these great extragalactic structures sit the voids. By volume, most of the universe is void—the filaments, walls, and clusters are tight agglomerations of galaxies, with high density but low volume. By contrast, the voids are almost completely empty, literally devoid of much matter at all. They're not totally empty, though, and detailed surveys of individual voids reveal a surprising feature.

If you zoom in on a portion of the cosmic web and look at the voids, you'll see a few dim dwarf galaxies. It's only by careful, detailed, prolonged observations that these galaxies reveal themselves to even the most dedicated astronomers, but they are there. You can play the same game that you would do for the entire cosmic web: map out their positions. And when you do that, you get a surprising answer—a faint echo of a cosmic web, embedded inside the voids.

The structure of the universe isn't quite a fractal, even though fractal cosmology has a long and storied history, but today for various reasons that word is the ultimate f-bomb in cosmological circles. But parts of the cosmic web do resemble a fractal, at least superficially; the cosmic web can be seen as a series of nested cosmic webs, each “level” dimmer and smaller than the last.³

This has only come to light—astronomy pun again, sorry—in the last few years. That's because astronomical surveys are set by brightness limits. You design a telescope, say, this big (I'm holding out my hands so you can see what I'm saying), with a sensor capable of detecting this many photons in each pixel. That sets a hard limit on the dimmest thing you can take a picture of.

In deep astronomy, objects can be dim because they're far away, or they can be dim because…they're dim. So the same technology that allows us to sweep farther into the distant reaches of the cosmos allows us to scan in detail the emptiest patches, looking for any signs of light.

It's only once you reach small enough scales—around five million parsecs—that the interactions between individual galaxies ruin the harmoniously nested picture. Also, once you reach large enough scales—around one hundred million parsecs—the cosmic web loses its distinctiveness. At those scales, the breakdown of cluster, filament, and void becomes meaningless. A single hundred-megaparsec patch of the universe looks pretty much the same as another hundred-megaparsec patch. Each patch will contain its own unique arrangements of structures, an intergalactic fingerprint pattern, but the statistics of those structures (like the number and sizes of voids, the typical distance between clusters, etc.) become the same.4

That is the scale of true cosmological uniformity. Remember the cosmological principle, the ground-base assumptions used in the mathematics of general relativity to describe and understand the history of the universe? Those mathematics obviously don't apply in the solar system or even the galaxy—there's too much other stuff going on, like cosmic rays and chemistry, that dominates the interactions at those scales.

But out past a hundred megaparsecs, all those complicated, messy, stubborn physics just blend together into a seamless blob, revealing the bland uniformity to satisfy the cosmological principle that our universe is homogeneous and isotropic. The cosmic web is huge and magnificent, a pattern stretching from one end of the universe to the other (ahem, the universe doesn't have an end, but the imaginative language was too juicy to pass up). The observable universe is about ninety billion light-years, or thirty gigaparsecs, across, and it's filled with this frightening, beautiful web of galaxies.

There's one more pattern in the cosmic web that I want to tell you about. It's a subtle one, but the detection of its very existence was an important—and recent—milestone in cosmology, and if we've learned anything in the hundreds of years of scientific astronomy, it's that subtlety is the name of the game.

Let's briefly zoom back to the early universe, before the decoupling of matter and radiation and subsequent unleashing of the cosmic microwave background. The universe was a hot, dense plasmatic soup of particles, crashing around violently in the tense bath of radiation that permeated the cosmos. In that primordial sauna, the first instabilities grew and would eventually lead to the grand structures that we observe today.

There were also sound waves. Yup, sound waves. Any medium, even a plasma, can support the existence of rippling waves of pressure, which is exactly what sound waves are. The early universe was a sonic cacophony of profound, booming, bone-quaking vibrations. The constant competition between matter and radiation for dominance in the young plasma created those sound waves, and they continually crashed throughout the universe.

That is, until matter and radiation went their separate ways. Once that divorce happened, the sound waves were left hanging, frozen in mid-plasma. A sound wave is a pressure wave, and the peak of the wave represents a region of higher-than-average density. Just like sound waves in air: when someone speaks, her larynx vibrates air molecules, pushing them together, creating a rippling effect of over-and-under densities that travel outward; eventually, those alternating densities of air molecules vibrate your eardrums.

Imagine if you could freeze a moment in time and see the sound wave hanging in the space between you and the speaker, with some air molecules smooshed closer together and some spread thinner. That was the state the atoms were left in just at the moment of recombination, 380,000 years into the history of the universe. It was a small effect, naturally, but it persisted, leading to a quiet but persistent bias in the grand arrangement of galaxies that would follow over the course of the upcoming billions of years.

This feature, known as baryon acoustic oscillations (baryon = normal matter; acoustic = like a sound wave; oscillations = wiggles), is detectable, quite clearly in fact, in the cosmic microwave background as large patchy blobs of temperature differences. And it's detectable, quite subtly, in the organization of galaxies at the very largest scales. Those frozen-in sound waves sit today at a scale of around 150 megaparsecs—an ever-so-slight “bump” in the average density of galaxies; a peculiar feature in the cosmic web; the largest structure in the universe carrying a birthmark from its primordial origins.5

I lied. The most fascinating thing about the large-scale structure of the universe, the arrangement of galaxies on the grandest of scales, isn't the fact that it resembles a weblike pattern; it's that it moves.

The cosmic web is alive. Not living-creature alive, but alive in the sense that it was different in the past and will be different in the future. The cosmic web has evolved, and its history is still being written. The pattern isn't static. Galaxies are buzzing around in orbits with the clusters. The filaments are really freight trains, ferrying loads of galaxies from the intergalactic rural countryside to the clusters. The walls are collapsing. The voids are growing.

Despite the frightening scales of the cosmic web, the physics that govern its evolution are ridiculously simple. It's just gravity. And time. Loads and loads of time.

The seeds were first laid down when the universe was less than a second old, and we've watched them grow like a nursery of baby chicks over the course of the first billion years, from random microscopic fluctuations to the first dense clumps, stars, galaxies, igniting in an explosion of energy at the cosmic dawn. But the process didn't end with the formation of the first lighthouses. Gravity does what gravity does, and slowly, over time, ever-larger structures arrived on the scene.

After the seeds were set, initially in a dark universe but then in a light-loving one, all you needed to do was wait a few billion years and voilà! Your very own cosmic web. Gravity and time—the recipe couldn't get simpler.

The iconic weblike pattern owes its existence to gravity, with its seeds laid down by the event of inflation itself. The same microscopic fluctuations that gave rise to the speckling of the cosmic microwave background eventually grew up—it's no wonder that I like to refer to that afterglow light pattern as the baby picture of the universe.

We live in what's called a hierarchical universe. The biggest objects—filaments, walls, and clusters—grew from gluing together smaller structures. As soon as galaxies assembled, they collapsed into distinct groups called, well, groups. The groups piled themselves into the walls and filaments and then streamed along those highways into the clusters to join their friends and their busy lives.6

But this process didn't happen all at once. Some galaxies have already completed their journey to their local cluster and have settled in nicely. Others are only recently getting a move on. Hence we get the weblike patchwork of in-process structure formation that we see in the universe today.

So it's fair to say that the cosmic web condensed out of the early, uniform primordial soup. Neat.

The best part is that we get to literally see this process, which played out at a glacial pace over the course of billions of years, frozen into the very pattern of galaxies that we observe with our deep surveys. That's because the science of astronomy is more closely related to paleontology than physics (please don't tell the astronomers I said that).

Light takes time to travel. It's fast, but not that fast, and especially not that fast compared to the cosmological distances it must travel to reach our eyes and optics. Our own sun is eight light-minutes away from the Earth. The photons that are making you squint or tanning your skin on a summer's day at the beach left the surface of the sun eight minutes ago. If the sun were to wink out of existence—a catastrophe, to be sure—we would at least have eight minutes of blissful ignorance before darkness fell upon us.

The measure of a light-year doesn't just mark out distances; it really does quantify a duration of time. A nameless star twinkling in the sky isn't just unfathomably distant from us in space, it's not present—the stars are neither here nor now. A couple thousand light-years to that speck of light means that we're seeing the star not as it is at this snap of my fingers but as it was thousands of years ago. Our image of the great Andromeda Galaxy isn't at the same age as the Milky Way.

As we probe ever deeper with our telescopes, we slowly reveal the younger universe. Surveys of the “nearby” universe (cosmologically speaking) exhibit the beautiful, intricate lacework patterns of the cosmic web. More distant galaxies trace out structures in an intermediate state of formation, and the faintest objects we can capture with our telescopes only hint at the beginnings of larger organizations. Past that are the dark ages, unlit by stars, which we are only beginning to probe. Once those are unveiled, we'll surely see a more primordial state (uh, at least I hope so).

At the very limit of our observations, we must switch to microwaves to map out the cosmic background radiation, the fossil relic from an infant cosmos.

The ancients viewed the universe from a geocentric point of view, with Earth at the center, surrounded by concentric spheres holding the moon, sun, and planets, surrounded by the uppermost celestial sphere hosting the not-so-distant stars themselves. In a strange quirk of perspectives, the time-capsule effect of light-travel time delays produces a picture of the universe with the Milky Way at the center, surrounded by a contemporary cosmic web, a hazier proto-web around that, a ring of blackness when the universe was plunged in darkness, and the thin shell of the primordial cosmic microwave background surrounding everything.

The universe has no center and no edge, but because light takes time to hop from its source to our telescopes, our view of the cosmos ends up recapitulating the models that we strove so hard to reject. I swear this is a coincidence—or at worst, a deep-seated psychological need to place our fragile humanity in some important position or focal point, lest the overwhelming largeness and complexity of the universe drive us mad.

Structures in our universe were built from the bottom up, from smaller objects to larger, grander cathedrals of light. Let's see how this process plays out in our local patch of the universe. To start, you should know that there's a place in our cosmos called the Great Attractor, and we're headed right for it.

Remember those maps of the ancient cosmic microwave background? And how tiny differences in the temperature from one spot to another reveal the face of the early universe? And how those differences pile up on each other to form the grandeur of the cosmic web? Good.

Before cosmologists could make that map, they had to subtract out our motion relative to the background itself. The Earth is orbiting the sun, the sun is orbiting the center of the Milky Way, and the Milky Way itself is buzzing through empty space. Altogether, it adds up to a brisk six hundred kilometers per second. Not too shabby.

This motion was first detected in the CMB itself, which is slightly (and “slightly” here means 0.0035 Kelvin) warmer on one side than the other due to the redshift of our motion. But we knew the speed of the Earth's orbit (thanks, Kepler!) and the speed of the sun's orbit in the great Milky Way. Add those together and it's…well, not six hundred kilometers per second. Ergo, the galaxy moves.

But if our home galaxy is moving, where is it moving to? Well, for one thing, we're on a collision course with the Andromeda Galaxy. That million-parsec space between our neighbor and us won't stay that comfortably large for long. In about five billion years—give or take a billion, depending on your exact definition of “collide”—we will collide with Andromeda, forming a single beefy galaxy yet to be named. Andromeda Way? Milkedromeda? The options are endless.

But that inevitable galactic headlong rush still isn't enough to explain the mystery of our intergalactic motion. Where exactly are we going?

As (bad) luck would have it, this is not an easy question to answer. For decades, astronomers have been carefully mapping the cosmos, starting with our local environs and steadily moving outward. But we can only make such measurements in galactically clear weather; the Milky Way is much wider around than it is thick, and it's chock-full of random blobs of gas and dust.

If you're an astronomer interested in blobs of gas and dust, then this is awesome. If you're a cosmologist, then this is beyond annoying. That space junk gets in the way of our pristine galactic images, making it difficult to find stuff through our Milky Way. This is especially true in the direction of our galactic core, the densest blob of gas and dust there is. It's a “zone” of the sky that we typically “avoid” because it's just too hard to look through.

It's a Zone of Avoidance.

And wouldn't you know it—the direction we're headed in, after you carefully subtract the motions of the Earth, sun, and Androky Way collision, is right there. Dang it.

We call it the Great Attractor, because who wouldn't?

The Milky Way, Andromeda, Triangulum, and a few dozen hanger-on dwarf galaxies form the Local Group, a gravitationally bound clump that's a handful of megaparsecs across.

The Local Group is a suburban town compared to the big city of the Virgo Cluster, more than 1300 galaxies crammed into a tight clump about a dozen megaparsecs down the road.

As I said before, clusters are the largest gravitationally bound objects in the universe, and you can probably guess where this is going: they're not done forming. Our Local Group is headed for the Virgo Cluster, leaving the country life for the glitz and glamour of the city.

But it doesn't stop there. Clusters themselves are gravitationally attracted to each other—they are massive balls of matter, after all—and are trying to form ever-larger structures. Unfortunately, definitions here get a little fuzzy. We know at the small end there are galaxies, groups, and clusters. And at the largest end is the cosmic web itself. In between is, well, stuff.

One name usually floated around is supercluster, which, just as it sounds, is something greater than a solo cluster. But despite finding them for decades, astronomers didn't really have a strict definition of one. If a structure was vaguely largish, it was a supercluster.

More recently we've come up with more robust definitions of these immense beasts. Since they're not gravitationally bound—they're not tied up into a tight little ball—we instead rely on the motions of galaxies to define them. Superclusters aren't yet done growing, so instead of looking at what they are, we must look at what they will be.

These new kinds of maps reveal the flows of our local universe—the agglomeration of structures into larger structures.7

Oh, and we've begun to map the universe within the Zone of Avoidance. While visible light gets snuffed out by our dusty galactic center, other wavelengths like radio and infrared sail right through, giving us the barest hint of the galaxies sitting on the far side of the Milky Way.

Combining the initial mappings within the Zone and a more respectable definition of a supercluster reveals what's going on with the Great Attractor. Remember how I said we live in a hierarchical universe? Yes, I'm repeating myself because it's that important.

Structures in our cosmos assemble themselves from smaller objects, and the Great Attractor is just the latest phase of that extragalactic construction project. The Local Group and a few other objects are collecting together toward the Virgo Cluster, which sits at the heart of the Virgo Supercluster. But the Virgo Supercluster is just a side branch of a much larger collection: the Laniakea Supercluster. And that supercluster has its own heart at the center: the Norma Cluster.

When you look in the direction of the Great Attractor, you're looking right at the Norma Cluster. But the Attractor isn't so much a thing as a place, the focal point of a process set in motion billions of years ago. Given enough time, every galaxy, group, and cluster within the Laniakea complex will collapse into a single uber-massive cluster.

Except it won't.

The cosmic web is not long for this universe. It's a transient, effervescent feature. A relic of the past, still persisting in the present but doomed in the future.

We live in a hierarchical universe, but we also live in an expanding universe. Structures like galaxies, groups, and clusters form despite this expansion. Although every galaxy is moving away from every other galaxy, this is only on average. Tiny instabilities in the early universe grow by gravity and, despite it being the wimpiest of the forces of nature, give it enough time and the pull of Newton and Einstein will do its indomitable work, growing structures bit by gaseous bit.

But the expansion is inevitable, and as I'm about to tell you in the next chapter, it's getting worse. Galaxies are still in motion, set off by gravitational interactions initiated in the distant past, but that motion is slowly grinding to a halt. Five billion years ago, the engines of creation shut down.

The Local Group will never reach the Virgo Cluster. The Virgo Supercluster will never reach Norma. Laniakea will never fully condense, and eventually it will be ripped apart.

In a few tens of billions of years, the cosmic web, with its beautiful, intricate lacework of filaments, walls, and knots—the largest pattern found in nature—will be gone.

All gone.