Glossary · Astrophysics
Cepheid Variable
A Cepheid is a pulsating star whose period is directly tied to its intrinsic luminosity. Discovered by Leavitt in 1912, this relation is the key to the cosmic distance scale.
Categorie
Étoile variable pulsante · Chandelle standard
Named After
Delta Cephei (Goodricke 1784)
Subclass Of
étoile variable intrinsèque pulsante
Relation Fondamentale
Période-luminosité (Leavitt 1912)
Periode Typique
1 à 100 jours (classiques) · 1-7 jours (type II)
Amplitude Typique
0,1 à 2 magnitudes
Luminosite
1 000 à 30 000 L☉ (Céphéides classiques)
Mecanisme
Kappa-mécanisme — ionisation/recombinaison de l'hélium doublement ionisé
Full definition
Imagine a star that beats like a heart, regularly swelling and contracting, its brightness rising and falling with each cycle. That is a Cepheid. And the icing on the cake: the slower the pulsation, the more luminous the star. Knowing the period means knowing intrinsic luminosity. Comparing intrinsic to apparent brightness gives distance — exactly the way you guess the distance of a known bulb from how bright it looks.
This is the most fruitful discovery of modern astronomy. In 1912, Henrietta Leavitt published the period-luminosity relation on 25 Cepheids in the Small Magellanic Cloud (all at roughly the same distance, so apparent-brightness differences reflected intrinsic ones). Edwin Hubble used this relation in 1923-1925 to measure the distance of the Andromeda 'nebula' (M31) via Cepheids he identified there: 285 kpc (modern value 780 kpc — off by a factor of 3, but the order of magnitude was enough to change the worldview). M31 was outside the Milky Way. Galaxies existed.
In 1929, Hubble published the distance-velocity relation: galaxies recede faster the farther they lie. That is the expansion of the Universe. Cepheids are at the heart of that discovery. To this day, they remain the 'first rung' of the cosmic distance ladder, calibrating type Ia supernovae which carry the measurement out to the edge of the observable Universe.
The physical mechanism is radial pulsation. In certain inner layers, helium ionizes and deionizes cyclically: doubly ionized, it becomes opaque, traps heat, the star heats up and expands; as it expands, helium cools and recombines, becomes transparent, heat escapes, the star contracts. The cycle resumes. This is the kappa mechanism, formalized by Arthur Eddington in 1917 and refined by Zhevakin and Cox in 1953-1963.
Cepheids occupy a specific region of the HR diagram called the 'instability strip' — there, and only there, the kappa mechanism can sustain regular pulsations.
The period-luminosity relation
The modern empirical formula in the V (visible) band:
M_V = -2.76 × log₁₀(P) - 1.40
where M_V is the absolute magnitude in V and P the period in days. A 10-day Cepheid has M_V ≈ -4.2 (~3,500 L☉), a 100-day Cepheid M_V ≈ -6.9 (~40,000 L☉).
Absolute calibration was long the Achilles' heel: you need distances to nearby Cepheids via direct parallax to calibrate the zero point. For decades, only a handful of galactic Cepheids had reliable parallaxes. The situation changed with Hubble (indirect parallaxes via Cepheids in clusters) and especially Gaia (ESA, 2013-), which measures astrometric distances to hundreds of Milky Way Cepheids to ~1 %. The SH0ES project (Adam Riess et al.) combines Gaia + Hubble + JWST to calibrate H₀ to ~1 % precision.
Two very different populations must be distinguished:
• Classical Cepheids (type I, DCEP). Young massive stars (4-20 M☉), metal-rich, in galactic disks. Period 1-100 days, L = 1,000-30,000 L☉. These are what Hubble and SH0ES use.
• Type II Cepheids (WVIR, BL Her). Old low-mass stars (~0.5 M☉), metal-poor, in globular clusters and halos. Period 1-7 days, ~4 mag less luminous than classical at equal period. Initially confused with classical — causing Hubble's factor-of-2 error in M31's distance, corrected by Baade in 1952.
Another related family: RR Lyrae, cousins of type II Cepheids, fainter (L ~ 50 L☉, P < 1 day), used for galactic distances.
Famous examples
Delta Cephei. The prototype. Apparent magnitude 3.5 to 4.4, period 5.366 days very precisely (so precise that the maximum can be predicted to the minute). 273 pc away. Naked-eye in Cepheus. Discovered variable by Goodricke in 1784.
Polaris (Alpha Ursae Minoris). YES, the Pole Star is a Cepheid! Period ~4 days, extremely small amplitude (~0.03 mag, imperceptible to the eye). Its pulsation even weakened in the 20th century then revived — subject of intense debate on its exact nature. The closest Cepheid to us (~130 pc).
Eta Aquilae. The brightest Northern-Hemisphere Cepheid, magnitude 3.5 to 4.4, period 7.177 days. Excellent summer beginner target.
Zeta Geminorum. Magnitude 3.6 to 4.2, period 10.15 days. Winter in Gemini.
RS Puppis. One of the most luminous known (L ~ 14,000 L☉), period 41.4 days. Famous for its dust halo, photographed by Hubble, which enabled a direct geometric distance measurement (1,970 pc) via light echoes.
V1 in Andromeda. The historical Cepheid used by Hubble in 1923 to prove that M31 is a separate galaxy. Magnitude 18.9 to 19.7, period 31.4 days. Cult object, honored with a commemorative plaque.
W Virginis. Prototype of type II Cepheids (WVIR). Period 17.27 days, old Population II star.
Cepheids in NGC 4258 and NGC 4424. Crucial for modern SH0ES calibration: M106 (NGC 4258) hosts a 22 GHz maser providing a direct geometric distance, which calibrates Cepheids observed there.
How do we observe them?
Naked eye or binoculars. Several galactic Cepheids are bright enough: Delta Cephei (mag 3.5-4.4), Eta Aquilae (mag 3.5-4.4), Zeta Geminorum (mag 3.6-4.2). Just compare them to neighbor stars of known magnitudes over several nights. AAVSO charts provide the comparison values. This is one of the most rewarding amateur-astronomy projects — you replay Leavitt's discovery live.
Amateur photometry. A cooled CCD or CMOS camera, free software (AstroImageJ) and a string of observing nights let you plot a complete light curve in a few weeks, at 0.01-0.02 mag precision. Measurements submitted to AAVSO feed databases used by professionals. Several French amateurs have contributed to published Cepheid catalogs.
Extragalactic Cepheid hunting. With 250 mm and a sensitive camera, you can image M31, M33, M81 and detect Cepheids in their spiral arms — with hour-long exposures or more. Demanding but achievable, and extremely satisfying: you literally measure the distance to another galaxy with your own data.
Professional missions. Gaia (ESA, 2013-) published in DR3 a catalog of more than 15,000 galactic Cepheids with parallaxes. Hubble and JWST follow Cepheids in 42 galaxies for the SH0ES H₀ calibration. The Roman telescope (NASA, launch planned 2027) will massively enlarge the sample.
Current stake: the 'Hubble tension'. The H₀ value from Cepheids + supernovae (SH0ES, H₀ ≈ 73.0 km/s/Mpc) differs by ~8 % from that from the cosmic microwave background (Planck, H₀ ≈ 67.4 km/s/Mpc). This ~5σ disagreement is one of the biggest current mysteries. Cepheids sit at the center of the debate. Our sky map tool lets you locate Delta Cephei, Eta Aquilae and several others.
Not to be confused with
Other pulsating variables. Not all pulsators are Cepheids. RR Lyrae (P < 1 day, old, less luminous) cover galactic distances. Delta Scuti (P ≈ hours, F-type main-sequence stars, small pulsations) are not standard candles. Miras (P ≈ 300 days, red giants, very luminous but not stable enough) serve as noisier candles. Cepheids occupy a specific niche of the instability strip.
Eclipsing binaries. Some eclipsing binaries may mimic a sinusoidal light curve, but eclipses are marked by flat bottoms and abrupt transitions at contact. Cepheids have an asymmetric light curve (rise faster than decline — the 'pulsation skew') that is continuous, without flat sections.
Classical Cepheids vs type II Cepheids. Big historical confusion. In 1929 Hubble did not know of the two populations. Baade (1952) corrected the error by showing that type II Cepheids are 4 magnitudes fainter than classical at equal period — M31's distance jumped from 250 kpc to ~700 kpc, and the age of the Universe doubled. Today we separate them by metallicity, galactic position, color and period.
Exoplanets via radial velocity. The Doppler shift of Cepheid spectral lines superficially resembles that of a star hosting a giant exoplanet. But a Cepheid physically pulsates (lines varying asymmetrically, up to tens of km/s), while an exoplanet induces a small, regular sinusoidal variation. Real traps have existed historically (e.g. 'exoplanet' around TrES-2 reinterpreted as stellar activity).
Pulsating white dwarfs (ZZ Ceti). Periods of hundreds of seconds, not days. Scales are wholly different — but the kappa-mechanism analogy is real (ionization of hydrogen or helium in the white dwarf).
Frequently asked
Why are Cepheids called 'standard candles'?
Because, knowing the pulsation period, you know the intrinsic luminosity — that's Leavitt's period-luminosity relation. If you know how much a light bulb really emits (absolute luminosity) and measure how much of it reaches you (apparent brightness), you can compute its distance. Cepheids are therefore cosmic bulbs of known power. They remain usable out to ~30 Mpc (100 million ly), beyond which the baton passes to type Ia supernovae (secondary standard candles).
Does a Cepheid pulsate forever?
No. A star is a Cepheid only during a specific phase of its evolution, when it crosses the instability strip on the HR diagram. For a typical ~5 M☉ star, this phase lasts a few million years — very short on astronomical timescales. Before, it was a quiet main-sequence star; after, it will continue its evolution as a red giant, then a planetary nebula (if initial mass < 8 M☉) or a supernova (above). Cepheid pulsation is therefore a transition episode, not a permanent state.
How do you measure a galaxy's distance using Cepheids?
You identify in the target galaxy variable stars with the characteristic Cepheid light curve, across images taken on several nights (then measure their period, typically 10-100 days for a nearby galaxy). The period-luminosity relation gives their absolute luminosity. Comparing with observed apparent brightness yields the distance modulus and therefore the distance in parsecs. Precision is a few percent. Hubble and JWST do this out to ~30 Mpc with 60-100-day Cepheids.
Which Cepheid is easiest to follow from your backyard?
Delta Cephei itself: magnitude 3.5-4.4, period 5.366 days, naked-eye in Cepheus (visible year-round from Europe, high in autumn). A comparison chart with neighboring Zeta and Epsilon Cephei is enough — just note Delta's magnitude each clear night. Within 3-4 weeks, you draw a beautiful light curve. Eta Aquilae in summer (magnitude 3.5-4.4, period 7.177 days, in Aquila near Altair) is another excellent pick.