This article traces the history of the dark matter concept in cosmology and astrophysics, from early anomalies in stellar and galactic observations through the development of the standard cosmological model and into the current era of experimental searches. For the main topic overview, see Dark Matter. For ongoing disputes about the nature and existence of dark matter, see Dark Matter - Debate. For the scientific consensus position, see dark-matter-consensus.

In 1844, Friedrich Bessel proposed that the irregular proper motion of Sirius and Procyon implied the presence of unseen companion stars. Though his inference was later confirmed by the discovery of white dwarf companions, this early reasoning - inferring invisible mass from gravitational effects on visible matter - established the methodological template that dark matter researchers would later employ at galactic scales.

In 1906, Henri Poincaré coined the term “matière obscure” (dark matter) in a discussion of Lord Kelvin's attempt to estimate the mass of the Milky Way by the velocities of stars. Kelvin had argued in 1904 that many stars might exist that are not self-luminous, and therefore not directly observable. Poincaré used the term descriptively, without developing it into a systematic hypothesis.

In 1922, Jacobus Kapteyn analyzed the vertical motions of stars near the solar neighborhood and estimated the local density of matter. He concluded that the dynamical mass was consistent with the visible mass, leaving little room for large quantities of unseen material - a conclusion that would later be revised.

In 1932, Jan Oort reanalyzed the vertical stellar kinematics in the solar neighborhood using improved data. Oort found a modest discrepancy suggesting the presence of some undetected matter in the Galactic plane, though his estimated deficit was small. He did not argue for a large-scale invisible mass component.

In 1933, Fritz Zwicky published an analysis of the Coma Cluster using the virial theorem - a relation between the kinetic and potential energy of a gravitationally bound system. Measuring the velocity dispersions of individual galaxies within the cluster and comparing them to the cluster's visible mass, Zwicky found that the visible matter was insufficient by a factor of roughly 400 to keep the cluster gravitationally bound. He called the inferred missing mass “dunkle Materie” (dark matter).

Zwicky's 1937 follow-up paper refined the estimate and proposed several candidate explanations, including faint stars, gas, and cold bodies. His finding attracted limited sustained attention for several decades, in part because the systematic uncertainties in early velocity measurements were large, and in part because galaxy cluster physics was not yet a well-developed field.

In 1939, Horace Babcock published measurements of the rotation curve of the Andromeda Galaxy (M31). He found that rotational velocities did not decline with distance from the galactic center as expected if most of the mass were concentrated in the luminous core - a Keplerian falloff - but instead remained roughly flat at large radii. Babcock attributed this to absorption effects or to a modified mass-to-light ratio in the outer disk, not to large-scale invisible matter.

In 1959, Louise Volders published rotation curve data for two galaxies, M33 and M101, also showing non-Keplerian behavior. Like Babcock, she did not interpret the result as evidence for a distinct dark matter component.

In the 1960s, Ken Freeman analyzed the surface brightness profiles and rotation curves of spiral galaxies. His 1970 paper noted that the observed rotation velocities in several galaxies implied more mass than the visible disk could account for, suggesting the existence of additional matter beyond the luminous region.

Beginning in the late 1960s, Vera Rubin and Kent Ford developed improved spectrographic techniques at the Carnegie Institution of Washington that allowed measurement of galactic rotation curves with greater precision and at larger radii than previously possible. Their 1970 paper on the rotation curve of M31 demonstrated that the curve remained flat far beyond the optical disk.

Over the following decade, Rubin, Ford, and collaborators published rotation curves for dozens of spiral galaxies of various morphological types and luminosities. By the late 1970s this body of work constituted a systematic observational result: flat rotation curves were a near-universal feature of spiral galaxies, not an artifact of individual measurements. The flat curves implied that mass continued to be distributed at large radii even where there was little or no luminous matter - a result widely interpreted as requiring a substantial non-luminous mass component, referred to as a dark matter “halo” surrounding each galaxy.

Rubin and Ford are frequently credited with establishing the dark matter problem as a central observational challenge for astrophysics, though historians of science have noted that their work built directly on Babcock, Volders, Freeman, and others.

In 1973, Jeremiah Ostriker and James Peebles published a theoretical analysis arguing that the stellar disks of spiral galaxies were dynamically unstable unless embedded in a massive spherical halo component. In 1974, Ostriker, Peebles, and Amos Yahil published a paper synthesizing dynamical evidence for unseen mass in and around galaxies and estimating the total mass of galactic halos.

In 1974, independently, Jaan Einasto, Ants Kaasik, and Enn Saar published a similar analysis from the Soviet side. These parallel publications brought the dark matter hypothesis to wider theoretical attention.

The nature of the dark matter - what it physically was - remained entirely open. Early candidates fell into two broad classes. Baryonic candidates included faint stars, brown dwarfs, cold gas, black holes, and compact remnants, collectively later termed MACHOs (Massive Astrophysical Compact Halo Objects). Non-baryonic candidates included hypothetical particles with weak or gravitational-only interactions.

In 1977, neutrinos were proposed as a dark matter candidate following the suggestion that they might carry a small but nonzero rest mass. This prompted detailed investigation of whether a neutrino-dominated universe could reproduce observed large-scale structure. By the early 1980s, simulations by Marc Davis and others showed that hot dark matter (fast-moving particles such as neutrinos) predicted structure on large scales inconsistent with observed galaxy clustering.

In 1982-1984, the Cold Dark Matter (CDM) model was developed in papers by Peebles, by Bond and Szalay, and particularly by George Blumenthal, Sandra Faber, Joel Primack, and Martin Rees. CDM proposed that dark matter consisted of slow-moving (cold), non-baryonic particles, produced the right large-scale structure, and became the dominant theoretical framework.

Big Bang nucleosynthesis (BBN) theory, developed in the 1940s and refined through the 1970s and 1980s, predicted the primordial abundances of light elements (hydrogen, helium, deuterium, lithium) as a function of the baryon density of the universe. Observations of these abundances consistently indicated a baryon density well below the total matter density inferred from dynamical and gravitational measurements. This implied that if dark matter existed in the quantities required to explain galactic rotation curves and cluster dynamics, the majority of it could not be ordinary baryonic matter - it had to be something else.

This constraint significantly narrowed the field of candidate dark matter particles and lent strong indirect support to non-baryonic candidates such as those arising from supersymmetric extensions of the Standard Model of particle physics.

Supersymmetric theories of particle physics, developed independently for reasons internal to high-energy physics, predicted the existence of a lightest stable supersymmetric partner particle. In the early 1980s, several groups - including Goldberg, Ellis, Pagels and Primack, and Steigman and Turner - recognized that this particle, if it existed, would have roughly the right properties and relic abundance from the early universe to constitute the dark matter. These particles were called WIMPs (Weakly Interacting Massive Particles).

The convergence of a particle physics prediction with a cosmological requirement - sometimes called the “WIMP miracle” - made WIMPs the dominant dark matter candidate through the 1980s and beyond, motivating a large experimental program of direct and indirect detection searches.

Einstein's general theory of relativity predicted that mass bends the path of light. In 1979, the first gravitational lens - a background quasar imaged into multiple components by a foreground galaxy - was observed. In subsequent years, gravitational lensing was developed as an independent method of measuring the mass distribution of galaxies and clusters.

By the late 1980s and 1990s, weak gravitational lensing surveys of galaxy clusters consistently showed that the total projected mass inferred from lensing exceeded the visible mass, corroborating the excess inferred from velocity dispersions. Strong lensing observations of clusters such as Abell 2218 produced detailed mass maps confirming extended invisible mass distributions.

In 1993-2000, the MACHO and EROS microlensing surveys searched for compact baryonic dark matter objects in the Milky Way halo by monitoring millions of stars in the Magellanic Clouds for the transient brightening caused by a foreground object passing in front of them. Both surveys found microlensing events, but at rates far below what would be required if MACHOs constituted all the dark matter, effectively ruling out a large fraction of the baryonic dark matter parameter space.

The Cosmic Microwave Background (CMB), first detected in 1965 by Penzias and Wilson, encodes information about the state of the universe roughly 380,000 years after the Big Bang. The angular power spectrum of temperature fluctuations in the CMB is sensitive to the amount and type of matter present. Measurements by the COBE satellite (1989-1993), the Boomerang and MAXIMA balloon experiments (2000), and especially the WMAP satellite (launched 2001) mapped the CMB with increasing precision.

The WMAP first-year results (2003) and subsequent data releases allowed precise determination of cosmological parameters. The fits to CMB data required two distinct non-luminous components: cold dark matter (contributing roughly 27% of the total energy density of the universe) and dark energy (roughly 68%). Ordinary baryonic matter accounted for only about 5%. These results, consistent with independent constraints from large-scale structure surveys and supernova distance measurements, established the Lambda-CDM (ΛCDM) model as the standard cosmological model.

In 2006, Douglas Clowe and collaborators published observations of the Bullet Cluster (1E 0657-56), a system consisting of two galaxy clusters that had undergone a high-speed collision. X-ray observations revealed that the hot gas (which contains most of the ordinary baryonic mass of a cluster) had been slowed and displaced by ram pressure during the collision. Weak gravitational lensing maps showed that the centers of mass, however, had passed through each other and remained near the galaxies rather than the gas.

The Bullet Cluster result was widely interpreted as strong evidence for a collisionless dark matter component - one that, like the galaxies, experienced little drag during the collision. It was cited as a significant challenge to alternative gravity theories that did not invoke dark matter.

The first purpose-built direct detection experiments for WIMPs were operated in the late 1980s. These experiments used shielded detectors underground to search for the nuclear recoil that would result from a WIMP scattering off an atomic nucleus.

Over subsequent decades, a series of increasingly sensitive experiments - including CDMS, XENON, LUX, PandaX, and others - have progressively lowered the upper limits on WIMP-nucleus interaction cross-sections without detecting a confirmed signal. The XENON1T experiment, reporting results in 2018, and its successor XENONnT, along with LUX-ZEPLIN (LZ), have constrained the WIMP parameter space by many orders of magnitude relative to original expectations.

The Large Hadron Collider (LHC) at CERN, which began operations in 2008 and reached design energy in 2015, searched for WIMP production via missing-energy signatures. No confirmed WIMP signal has been reported.

In 1983, Mordehai Milgrom published a phenomenological proposal called Modified Newtonian Dynamics (MOND), which posited that Newton's second law required modification at very low accelerations. MOND was motivated by the observation that flat rotation curves appeared when and only when the centripetal acceleration fell below a universal threshold value. In a subset of galactic rotation curve cases, MOND reproduced observations without invoking dark matter.

Subsequent relativistic extensions of MOND, including TeVeS (Tensor-Vector-Scalar gravity) proposed by Jacob Bekenstein in 2004, aimed to produce a covariant theory that could be confronted with lensing and cosmological data.

Whether the flat rotation curves of spiral galaxies observed by Rubin, Ford, and others in the 1970s uniquely require a dark matter halo, or are consistent with modified gravity frameworks such as MOND, remains contested; see dark-matter-controversy-rotation-curves.

The interpretation of the Bullet Cluster as a decisive refutation of MOND-type theories is disputed by some researchers who argue that alternative gravity theories can accommodate the observations or that the collision parameters themselves require explanation; see dark-matter-controversy-bullet-cluster.

Whether the failure of WIMP direct detection experiments through the 2010s and 2020s constitutes evidence against dark matter as a whole, or merely constrains a specific class of candidates while leaving the broader hypothesis intact, is a matter of ongoing debate; see Dark Matter - Debate.

The historical priority question of who first identified the dark matter problem in a modern sense - with various claims made for Zwicky, Oort, Babcock, or Rubin and Ford - is discussed by historians of science; see dark-matter-controversy-historical-priority.

Whether anomalies in small-scale structure (the “missing satellites problem,” “cusp-core problem,” and “too-big-to-fail problem”) represent observational selection effects, baryonic feedback processes, or fundamental challenges to cold dark matter, remains unresolved; see dark-matter-controversy-small-scale-structure.

1. Friedrich Wilhelm Bessel, “Ueber Veränderlichkeit der eigenen Bewegungen der Fixsterne,” Astronomische Nachrichten 22 (1844): 145-160.

2. Lord Kelvin (William Thomson), “On the Motions of Monochromatic Light in a Homogeneous Medium,” in Baltimore Lectures on Molecular Dynamics and the Wave Theory of Light (London: C.J. Clay and Sons, 1904), 274-281.

3. Henri Poincaré, “La Voie Lactée et la théorie des gaz,” Bulletin de la Société Astronomique de France 20 (1906): 153-165.

4. Jacobus C. Kapteyn, “First Attempt at a Theory of the Arrangement and Motion of the Sidereal System,” Astrophysical Journal 55 (1922): 302-328.

5. Jan H. Oort, “The Force Exerted by the Stellar System in the Direction Perpendicular to the Galactic Plane and Some Related Problems,” Bulletin of the Astronomical Institutes of the Netherlands 6 (1932): 249-287.

6. Fritz Zwicky, “Die Rotverschiebung von extragalaktischen Nebeln,” Helvetica Physica Acta 6 (1933): 110-127.

7. Fritz Zwicky, “On the Masses of Nebulae and of Clusters of Nebulae,” Astrophysical Journal 86 (1937): 217-246.

8. Horace W. Babcock, “The Rotation of the Andromeda Nebula,” Lick Observatory Bulletin 19 (1939): 41-51.

9. Louise Volders, “Neutral Hydrogen in M 33 and M 101,” Bulletin of the Astronomical Institutes of the Netherlands 14 (1959): 323-334.

10. K.C. Freeman, “On the Disks of Spiral and S0 Galaxies,” Astrophysical Journal 160 (1970): 811-830.

11. Vera C. Rubin and W. Kent Ford Jr., “Rotation of the Andromeda Nebula from a Spectroscopic Survey of Emission Regions,” Astrophysical Journal 159 (1970): 379-403.

12. Jeremiah P. Ostriker and James E. Peebles, “A Numerical Study of the Stability of Flattened Galaxies: Or, Can Cold Galaxies Survive?” Astrophysical Journal 186 (1973): 467-480.

13. Jeremiah P. Ostriker, James E. Peebles, and Amos Yahil, “The Size and Mass of Galaxies, and the Mass of the Universe,” Astrophysical Journal Letters 193 (1974): L1-L4.

14. Jaan Einasto, Ants Kaasik, and Enn Saar, “Dynamic Evidence on Massive Coronas of Galaxies,” Nature 250 (1974): 309-310.

15. George R. Blumenthal, Sandra M. Faber, Joel R. Primack, and Martin J. Rees, “Formation of Galaxies and Large-Scale Structure with Cold Dark Matter,” Nature 311 (1984): 517-525.

16. C. Alcock et al. (MACHO Collaboration), “The MACHO Project: Microlensing Results from 5.7 Years of Large Magellanic Cloud Observations,” Astrophysical Journal 542 (2000): 281-307.

17. P. Tisserand et al. (EROS-2 Collaboration), “Limits on the Macho Content of the Galactic Halo from the EROS-2 Survey of the Magellanic Clouds,” Astronomy and Astrophysics 469 (2007): 387-404.

18. C.L. Bennett et al. (WMAP Science Team), “First-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Preliminary Maps and Basic Results,” Astrophysical Journal Supplement Series 148 (2003): 1-27.

19. Douglas Clowe et al., “A Direct Empirical Proof of the Existence of Dark Matter,” Astrophysical Journal Letters 648 (2006): L109-L113.

20. Mordehai Milgrom, “A Modification of the Newtonian Dynamics as a Possible Alternative to the Hidden Mass Hypothesis,” Astrophysical Journal 270 (1983): 365-370.

21. Jacob D. Bekenstein, “Relativistic Gravitation Theory for the Modified Newtonian Dynamics Paradigm,” Physical Review D 70 (2004): 083509.

22. E. Aprile et al. (XENON Collaboration), “Dark Matter Search Results from a One Tonne × Year Exposure of XENON1T,” Physical Review Letters 121 (2018): 111302.