A galaxy can be “weighed” in two ways: estimate the mass of its stars and gas, or ask how much gravity their motion requires. For many systems, these estimates disagree. More intriguingly, a similar mass shortfall appears through bent light and the history of cosmic structures.
Dark matter names an additional matter component inferred from these effects in the standard gravitational picture, although it is not visible as ordinary light. The evidence for its gravitational effects is much better established than the identity of the particles it might contain. PDG: the case for dark matter, §27.1.
Motion reveals how much mass is needed
Astronomers measure spectral shifts in light from stars and gas, account for a galaxy’s inclination and obtain a rotation curve: orbital speed as a function of distance from the centre. This is not a photograph of an entire orbit. Separately, they construct an ordinary-matter model from the distribution of starlight and gas.
In many disk galaxies, speed remains high and approximately constant well beyond the bright centre, while the accounted-for stars and gas provide insufficient gravity. The standard model explains this with an extended dark-matter halo. Not every galaxy has the same rotation curve; uncertainties in distance, inclination and stellar mass genuinely matter. SPARC: measured rotation curves and stellar and gas disk models.
Saying “dark matter holds a galaxy together” is insufficient. The specific test is whether the chosen mass reproduces velocities at different radii. Counting only bright stars while forgetting gas or dim ordinary material cannot supply the answer. PDG: ordinary matter in dynamical estimates, §27.4.1.
Branch · With a formulaOptional · you can skip this
What constant speed at a larger radius implies
For an illustrative circular orbit in the Newtonian field of a spherical mass distribution:
M( < r) = (r(v)^(2))/(G).
M( < r) is mass inside radius r, v is circular speed and G is the gravitational constant. Metres multiplied by squared metres per second, then divided by G, give kilograms.
If hypothetically r_(2) = 4r_(1) and v_(2) = v_(1), then M( < r_(2)) = 4M( < r_(1)). The outer region is still adding mass. If almost all the mass were already inside r_(1), speed at 4r_(1) would instead halve. Equal radii and speeds give a mass ratio of one.
These are two idealized predictions, not data for a particular galaxy. A disk is not a sphere: precise stellar and gas contributions require disk geometry, rather than unqualified use of this formula. OpenStax: circular orbits, §13.4, equation 13.7.
Light provides a different way to weigh
Gravity changes light’s path. Images of distant galaxies behind a massive cluster are distorted, sometimes forming arcs or multiple images. This is gravitational lensing. Source shapes, positions and distances can be used to reconstruct the gravitating mass distribution.
This weighing method does not require light from the dark component itself. It does require a lens model: a picture with a coloured “dark cloud” usually combines observations with a reconstructed mass map. The map’s colour is not the colour of unknown particles. NASA: mapping mass from distorted light.
A revealing case is the collision of galaxy clusters in the Bullet Cluster. Hot gas, containing most of these clusters’ ordinary matter, interacted and slowed. The main lensing-inferred mass concentrations lie closer to the galaxies than to the principal gas peaks. This strongly supports an additional component that behaves differently from gas; the map does not identify its particle. Clowe et al., 2006: original observations and analysis.
The figure compares three different lines of evidence, each with its own measurements and assumptions.
One hypothesis, different tests
Motion in galaxies
- Observe
- Star and gas velocities at different distances from the centre.
- Infer with a model
- The gravitational field and mass distribution required by that motion.
- Account for
- Inclination, distance, stellar and gas mass, and the system geometry.
Bending light
- Observe
- Shapes and positions of images of distant sources.
- Infer with a model
- The gravitating mass distribution along the line of sight.
- Account for
- Lens geometry, distances and projection; not particle light.
The early and late cosmos
- Observe
- The statistical pattern of relic light and galaxy clustering.
- Infer with a model
- Composition and structure growth in a chosen cosmological model.
- Account for
- Initial conditions, parameters, ordinary matter and gravity laws.
The explanation must work in both the early and late cosmos
The cosmic microwave background—ancient light now observed at microwave wavelengths—retains small temperature contrasts from the early universe. Its statistical pattern and later galaxy clustering test cosmic composition and structure growth. In standard cosmology, an additional component that builds gravitational irregularities early and couples only weakly to the luminous plasma helps reconcile these epochs. “Cold” dark matter means motion slow enough for this structure formation, rather than an ordinary thermometer reading in kelvin. PDG: structure and dark matter, §27.1.
In the baseline ΛCDM model, dark matter accounts for roughly 85% of all matter, not 85% of the universe’s total contents including dark energy. This is an inference from observations within a model. Ordinary material outweighs bright stars alone, but hiding all the required mass in gas and dim objects does not work: primordial elements and the microwave background also constrain the amount of ordinary atomic constituents. PDG: the dark-matter fraction, §27.1; PDG: separate estimates of baryonic matter, §24.4.
Modified gravity laws can also be tested. Some describe particular rotation regularities successfully, but an explanation must also handle lensing, colliding clusters and cosmological observations. Success with one curve does not replace that combined test. Dark-matter models also retain questions about small-scale distributions and the effects of ordinary gas and star formation. PDG: alternatives and small scales, §27.1, §27.4.2.
Weighing the invisible does not identify its ingredients
WIMPs, axions and other proposed particles are candidates, rather than names of an established substance. Laboratory detectors seek rare interactions with known matter; astronomical searches look for possible transformation products. Any such inference must distinguish the required signal from ordinary backgrounds. PDG: models and direct searches, §27.5–27.6.
Branch · A dated detailOptional · you can skip this
September 2026: an unusual event is not yet a discovery
On 1 September 2026, LUX-ZEPLIN reported one event in its xenon detector in a region where few events were expected from known backgrounds. It is an interesting candidate to investigate, but the collaboration explicitly did not claim a dark-matter discovery. More data and background checks are needed. LZ: the official result; the original paper, arXiv 2609.02823.
The announcement date is not the interaction date. Nor does this result turn every previous limit into a particle detection. This detail was checked on 30 October 2026: the original work on arXiv is a preprint; the preliminary nature of the result must accompany the report.
Dark matter therefore explains a connected set of gravitational observations while leaving its physical composition open. Confidence in an additional gravitating component and confidence in a particular candidate are different things. To compare mass and gravity, revisit mass and weight.