On the night of October 5, 1923, Edwin Hubble, using the 100-inch Hooker Telescope at Mount Wilson Observatory in California — then the largest in the world — captured a photographic plate of the Andromeda Nebula. Examining it the next day, he identified a Cepheid variable in the nebula's outskirts. Applying Leavitt's period-luminosity relation, the distance came out to ~285 kpc — later refined to ~770 kpc, about 2.5 million light-years. Andromeda was not part of the Milky Way; it was another galaxy — an island universe, as Kant had speculatively called such things in 1755. The Great Debate of 1920 between Harlow Shapley, who argued the Milky Way was the universe, and Heber Curtis, who argued the spiral nebulae were external galaxies, was settled by Hubble's plate. By 1929 Hubble had also discovered the redshift-distance relation — Hubble's law — with the implication that the universe is expanding.
The Hubble morphological classification (1926) divides galaxies into ellipticals (smooth spheroidal, old stars, little gas), spirals (disk + bulge + arms, ongoing star formation), barred spirals (the Milky Way's class), lenticulars, and irregulars. The Milky Way is a barred spiral roughly 100,000 light-years across containing about 10¹¹ stars; the Sun sits in the disk, ~25,000 light-years from the galactic centre. The galactic centre hosts Sgr A*, a supermassive black hole of ~4.1 × 10⁶ solar masses, whose mass was measured precisely by tracking stellar orbits in the central parsec — a thirty-year programme that earned Andrea Ghez and Reinhard Genzel the 2020 Nobel Prize. Every studied galaxy with a bulge has a supermassive black hole at its centre with mass ~0.5% of the bulge mass — the M-σ relation — implying coordinated formation between SMBHs and bulges through some still-contested mechanism. Galaxies cluster gravitationally into groups ranging from the Local Group (about eighty galaxies, dominated by Andromeda and the Milky Way) through the Virgo Cluster (~1,300 galaxies, 16 Mpc away) to the Coma Cluster (~1,000 galaxies, 100 Mpc away). Clusters are filled with hot X-ray-emitting plasma — the intracluster medium at 10⁷ to 10⁸ K. At larger scales, galaxies trace the cosmic web, clustering along filaments with voids between them. The most consequential fact is that most of the gravitating mass is invisible: spiral-galaxy rotation curves stay flat far beyond the visible disk, gravitational lensing in clusters is several times what visible matter would produce, and cosmological simulations need cold dark matter to reproduce the observed structure. The dark-matter halo of any large galaxy holds five to ten times the visible mass; identifying what particle dark matter consists of is one of the central open problems in physics.
JWST has been observing galaxies at extreme redshifts (z > 10, when the universe was less than 500 million years old). Early 2022–2023 results suggested galaxies more massive and numerous than expected, producing a brief 'JWST broke cosmology' narrative; more careful 2023–2024 analyses showed many high-z candidates were photometric-redshift artefacts. The settled finding is that the early universe produced more and brighter galaxies than pre-2022 models predicted, but not in a way that breaks ΛCDM cosmology. The Event Horizon Telescope produced the first imaged supermassive-black-hole shadows — M87* in April 2019 and Sgr A* in May 2022 — both consistent with general-relativity predictions for Kerr black holes. The Vera Rubin Observatory will, over its ten-year survey, produce time-domain data on every galaxy in the visible southern sky every few nights — a transformative dataset for tidal-disruption events, supernovae, and gravitational-lens monitoring.