Galaxies
The Scale at Which Stars Live
What a Galaxy Is
A galaxy is what stars do when there are billions of them held together by gravity. It is the natural scale of organization above the single star and below the cosmic web. The Milky Way you live in is one galaxy. The fuzzy patch of light in the constellation Andromeda is another galaxy. Across the observable universe, the current best estimate is that there are something like a few hundred billion galaxies, give or take, each one a self-contained gravitational system with its own population of stars, gas, dust, and dark matter. Most large galaxies also harbor a supermassive black hole at the center, though many dwarf galaxies do not.
The realization that galaxies are separate "island universes" outside the Milky Way is shockingly recent. Until the mid-1920s, even professional astronomers debated whether the smudges of light visible in deep-sky photographs were nearby nebulae of gas inside our own galaxy or distant galaxies in their own right. Edwin Hubble settled the question in 1924, having identified individual Cepheid variable stars inside the Andromeda nebula the previous year and used their known brightness pattern to calculate the distance. Andromeda was far too far away to be part of the Milky Way. It was its own galaxy, and so were the rest of the smudges. Almost overnight, the size of the known universe grew by many orders of magnitude.
A Family of Shapes
Galaxies do not all look the same. Hubble classified them in 1926 into three broad families that astronomers still use today. Spirals are flat rotating disks with visible arms wrapping outward from a central bulge - the Milky Way is one of these. Ellipticals are puffy three-dimensional balls of stars with little internal rotation and little ongoing star formation. Lenticulars sit in between, with a disk but no spiral arms. A small fraction are irregular - galaxies that have been bent out of shape by an interaction with a neighbor.
The Hubble classification was originally proposed as a kind of evolutionary sequence, with galaxies starting as one type and progressing to another. That interpretation is now known to be wrong. A given galaxy mostly stays the type it formed as, unless something dramatic happens to it - a merger with a comparable-sized neighbor, for example, can scramble a pair of spirals into a single elliptical. The morphology is the result of a galaxy's history, especially how it grew and how violently it interacted with others. The classification itself has aged remarkably well, even as the interpretation has changed.
Our Own Galaxy
The Milky Way is a barred spiral galaxy. From outside, if you could see it from above, you would see a roughly disk-shaped collection of stars with a brighter central bulge crossed by an elongated bar, and prominent spiral arms wrapping outward from the bar's ends. The disk is around one hundred thousand light-years across and only a few thousand light-years thick - a remarkably flat structure. It contains somewhere between one hundred and four hundred billion stars; the uncertainty comes from how many faint low-mass stars hide in the disk and how much of the dark interior we can actually see.
The Sun sits roughly halfway out from the center, on a small structure called the Orion Spur, between two larger arms. It takes the Sun roughly two hundred and thirty million years to complete one orbit around the galactic center. The galactic center itself contains a supermassive black hole called Sagittarius A-star, with a mass around four million times that of the Sun. The Event Horizon Telescope produced the first direct image of Sgr A-star's shadow in 2022, confirming the picture astronomers had built up over decades by watching individual stars orbit the central darkness at thousands of kilometers per second.
Outside the visible disk, the Milky Way sits in a much larger, much rounder cloud of dark matter that extends far beyond where any stars or gas are visible. That halo is what holds the galaxy together at large radii. The visible matter is only a few percent of the total gravitational mass of the system. Almost everything you can see inside the disk is a thin film of stars riding on top of an invisible reservoir.
Built From the Top Down by Dark Matter
The standard picture of galaxy formation starts long before any stars existed. In the early universe, density fluctuations imprinted at the time of inflation grew through gravity into a network of slightly over-dense and slightly under-dense regions. Dark matter, which does not feel pressure from light or thermal radiation, was free to collapse into structure earliest. By the time the universe became transparent about 380,000 years after the Big Bang, dark matter had already formed a clumpy network of halos.
After recombination, ordinary gas began to feel the gravitational pull of these dark halos and fell inward. As the gas fell, it heated up. Once it had concentrated enough in the center of a halo, it could begin to radiate away its heat through atomic transitions, cool down, and settle into a denser, flatter rotating structure. When the gas got cold and dense enough, it fragmented into the first stars. The first galaxies formed inside the first dark halos within a few hundred million years of the Big Bang.
Small galaxies formed first, then merged into progressively larger ones over time. This hierarchical assembly picture is what computer simulations have predicted for several decades, and it broadly matches what telescopes see at all distances - which means looking at all past epochs. Smaller galaxies are visible at earlier times. Larger galaxies are visible at later times, including most of the present-day giants in our cosmic neighborhood.
JWST and the Tension at Early Times
The James Webb Space Telescope, operating since 2022, has been able to image galaxies as they existed when the universe was only a few hundred million years old. The first such observations produced a surprise. Some of the earliest galaxies looked unexpectedly bright and unexpectedly massive - far brighter than what the standard hierarchical-assembly picture predicted for that era. For a year or two, this looked like a serious problem. Headlines suggested cosmology might be in trouble.
The picture has cooled considerably since then. Follow-up spectroscopic measurements at the LBT, ALMA, and JWST itself have refined the mass estimates of many of these early galaxies, and most of them now look less extreme than the initial broad-band photometry suggested. The galaxies are real and they really are at extremely high redshifts; they are just not as massive as the first quick estimates implied. The dust content, star-formation efficiency, and the contribution of accreting black holes to the brightness all turned out to be different from the simplest assumptions.
A few genuinely puzzling galaxies remain. The most massive systems at the earliest times still sit on the high end of what current models predict, and the so-called "little red dots" - extremely compact, red, point-like sources at high redshift - are an actively debated population whose nature is not yet settled. They may be small galaxies hosting overweight central black holes, or compact starbursts surrounded by dust. Through 2026 the field has moved from "crisis" to "active tension." Cosmology has not had to change, but the story of how galaxies grow at early times needs more detail.
When Galaxies Collide
Stars inside a galaxy are extraordinarily far apart relative to their size. Two galaxies passing through each other almost never have stars physically collide. What does collide is the gas and the gravitational fields. The interaction can tear long streamers of stars and gas off both galaxies, redirect them onto new orbits, and over hundreds of millions of years cause the two galaxies to spiral together into a single merged object.
Our nearest large galactic neighbor, Andromeda, is currently approaching the Milky Way at roughly one hundred kilometers per second. For a century the textbook prediction has been a merger: in about four to five billion years the two galaxies begin a first close pass, then over the following billion or so years swing past, return, and merge into a single larger system. Recent analyses that fold in the Large Magellanic Cloud and sharper Gaia measurements have softened that to roughly a coin-flip over the next ten billion years, so a merger is now best described as likely rather than guaranteed. By the time the merger completes, the Sun will already be well into its red giant phase, and the Earth's future will be grim independently of what the galaxies are doing. None of this is a danger to anyone alive today, and the merger itself, once it happens, will look spectacular but pose almost no danger to any individual star within either galaxy. Stars are too far apart to crash. The merger is a gravitational reshuffling, not a destructive collision.
Galaxy mergers play an important role in galaxy evolution. Most large elliptical galaxies are believed to have formed through mergers of disks. Each major merger pours fresh gas onto the central supermassive black hole, briefly lighting it up as an active galactic nucleus before the fuel is exhausted again.
When the Central Black Hole Lights Up
Every massive galaxy is thought to have a supermassive black hole at its center, with a mass anywhere from about a million times the Sun's mass to ten billion times. Most of the time, these central black holes are quiet - they are not actively feeding, and their presence is only detectable through the orbits of stars and gas in their immediate vicinity. Occasionally, gas falls onto a central black hole at a high rate. As it spirals in, it heats up to extreme temperatures and emits enormous amounts of radiation. The galaxy core can outshine the rest of the galaxy combined. This is called an active galactic nucleus, or AGN.
The most luminous AGN are called quasars. A bright quasar can outshine its host galaxy by a factor of a hundred. Some AGN also launch narrow, fast jets of charged particles perpendicular to the accretion disk. These jets travel at nearly the speed of light and can extend many times the diameter of their host galaxy. The energy released by AGN activity is so large that it can push gas out of the galaxy entirely, slowing down or stopping star formation in the host. This "AGN feedback" is now considered an essential part of how the largest galaxies grow and stop growing - without it, simulations produce galaxies that are too massive compared to what we actually see.
By the mid-2020s, evidence has accumulated that supermassive black holes and their host galaxies co-evolve. The mass of the central black hole correlates tightly with the mass of the bulge of the host galaxy across a wide range of sizes, suggesting that the two grow in lockstep. How that lockstep is enforced - whether through feedback, through coincident mergers, or through a more subtle joint history - is one of the most actively studied questions in galaxy evolution today.
The First Solid Evidence for Dark Matter
If a spiral galaxy were held together only by the gravity of its visible stars and gas, the orbital speed of stars at increasing distances from the center would gradually decline, much like the orbital speeds of planets fall off with distance from the Sun. The faster speeds belong closer in, where most of the gravitational pull is concentrated.
In the 1970s, Vera Rubin and Kent Ford measured the rotation curves of dozens of spiral galaxies in detail and found something striking. The orbital speeds did not fall off with distance. They stayed roughly constant, all the way out to the edges of the visible disks, and in some cases even beyond. The only way this can happen is if there is much more mass present than the visible stars and gas suggest, with most of that extra mass distributed in a roughly spherical halo extending well past the visible galaxy. The extra mass does not emit or absorb light, and it does not interact via the strong or electromagnetic forces. It is gravitationally there but optically invisible. Dark matter.
Rotation curves were among the cleanest pieces of evidence for dark matter, and they remain among the strongest. Later evidence from gravitational lensing, the cosmic microwave background, large-scale structure, and the dynamics of galaxy clusters has all pointed in the same direction. Most of the gravitational mass of the universe is in a form that does not interact with light. What it actually is - a new species of particle, something more exotic, or a modified theory of gravity disguised as dark matter - remains the largest unanswered question in physics. Galaxy rotation curves are where that question first sharpened into experimental fact.
Galaxies Belong to Groups
Galaxies are not distributed randomly through space. They cluster. The Milky Way belongs to a small system called the Local Group, which contains a few dozen galaxies including Andromeda and a population of much smaller dwarf galaxies. The Local Group itself lies on the outskirts of a much larger structure called the Virgo Supercluster, which spans about a hundred million light-years and contains thousands of galaxies. Larger still are the cosmic-web filaments and voids that connect superclusters into the largest known structure in the universe.
Rich galaxy clusters, like Coma and Virgo, contain hundreds to thousands of galaxies bound by mutual gravity. Their centers are typically dominated by enormous elliptical galaxies that grew by merging with their neighbors over billions of years. Between the cluster galaxies sits a thin but very hot gas heated to tens of millions of degrees, visible in X-rays. This intracluster gas alone contains more total ordinary mass than all the stars in the cluster combined. Wrapped around the whole system is a dark matter halo larger than the visible cluster, holding everything together. By measuring how gravitational lensing distorts the images of more distant galaxies behind a cluster, astronomers can map the total mass distribution, including the dark matter. The total mass of a large cluster can exceed a thousand trillion solar masses, with stars contributing only a few percent of it.
The Bigger Picture
Galaxies are the bridge between particle physics and cosmology. They are where dark matter shows up most cleanly. They are where supermassive black holes assemble. They are where stars form, evolve, die, and seed the next generation of stars with the elements they made. They are where every observed structural complexity in the universe sits - from the disks holding planetary systems to the clusters holding hundreds of galaxies. Almost everything physics studies above the scale of a single star sits inside a galaxy somewhere.
Through 2026, the field has a clear standard model of galaxy formation - hierarchical assembly inside dark-matter halos, regulated by feedback from supernovae and active galactic nuclei. The model is supported by an enormous body of observation and simulation. It also leaves real, specific tensions unresolved: the abundance of very massive galaxies at very early times, the smallest-scale dark-matter predictions versus what dwarf galaxies actually show, and the precise way in which black holes and their hosts co-evolve. None of these tensions are large enough to overturn the standard picture. All of them are large enough to keep the next generation of telescopes and simulations busy. Galaxies have been a productive object of study for a century. They have not stopped being one yet.
