Mass-Luminosity Relation

The mass–luminosity relation is the single most productive regularity in stellar astrophysics: across the main sequence, a star's luminosity is an extraordinarily steep power of its mass, $L \propto M^{\alpha}$ with $\alpha$ near $3.5$ . Double a star's mass and it shines roughly eleven times brighter; a $10\,M_\odot$ star outshines the Sun three thousandfold. Since fuel supply grows only linearly with mass, this steepness is what makes the most massive O stars burn out in a few million years while the feeblest M dwarfs will still be shining long after the present Universe has aged beyond recognition. The relation wears two faces. Empirically it was read off from detached eclipsing and visual binaries — the only systems where a mass can be measured directly, from Kepler's third law and a luminosity from photometry plus a parallax. Theoretically it falls out of the coupled equations of stellar structure with almost no input from nuclear physics: hydrostatic equilibrium fixes how pressure and temperature run with depth, radiative diffusion fixes how fast heat escapes, and the ferocious temperature sensitivity of fusion merely pins the star at the temperature that supplies exactly the luminosity the structure demands. The exponent is not a constant of nature — near a solar mass it steepens toward $4$ , for very massive stars it flattens toward $1$ as the Eddington limit is approached, and below $0.4\,M_\odot$ the fully convective dwarfs change slope again. Downstream, the relation feeds stellar lifetimes, the dating of clusters, the reading of the Hertzsprung–Russell diagram, and the mass-to-light ratios used to weigh galaxies and expose their dark matter.

Key takeaways

On the main sequence, luminosity is a steep power of mass, $L \propto M^{\alpha}$ with $\alpha \approx 3.5$ on average — a factor-of-two in mass is roughly an order of magnitude in light.
It emerges from hydrostatic equilibrium plus radiative transport: the power leaking out of a self-gravitating gas ball is set by gravity and opacity, not by the fusion rate.
A homology estimate at constant opacity gives $L \propto \mu^{4} M^{3}/\kappa$ ; Kramers opacity steepens it toward $\alpha \approx 4$ , the radiation-dominated Eddington limit flattens it toward $L \propto M$ .
Because fuel grows as $M$ but burn rate as $L \propto M^{\alpha}$ , lifetime scales as $t_{\rm MS} \propto M^{1-\alpha} \approx M^{-2.5}$ — massive stars are brilliant and doomed.
Calibrated on binary-star masses, it is the backbone for weighing stars from their light, dating clusters by their turnoff, and anchoring the stellar mass-to-light ratio.
Key idea: for a radiative, gas-pressure main-sequence star, hydrostatic equilibrium plus radiative diffusion force $L \propto \mu^{4} M^{3}/\kappa$ almost independently of the nuclear physics; fusion then merely sets the radius and surface temperature, so luminosity is governed by how gravity compresses the gas and how photons diffuse out.

The core insight

For a radiative, gas-pressure main-sequence star, requiring hydrostatic equilibrium and radiative energy transport forces $L \propto \mu^{4} M^{3}/\kappa$ almost independently of the nuclear physics; the reaction rates then merely fix the radius and surface temperature. Hence luminosity is governed by how gravity compresses the gas and how photons diffuse out, not by the details of fusion.

Where this lives today

The mass–luminosity relation is calibrated and used harder now than ever. Modern detached eclipsing binaries pin stellar masses and radii to better than one or two percent, tightening the empirical MLR band by band, while Gaia's astrometry extends dynamical masses to the faint low-mass dwarfs where the relation steepens and models have been shakiest, and asteroseismology cross-checks masses from oscillation frequencies without a binary at all. Downstream, integrating the MLR over an initial mass function yields the stellar mass-to-light ratio of a population; comparing that stellar $M/L$ with the dynamical $M/L$ from rotation curves and velocity dispersions (Binney & Tremaine) is a primary route to the dark matter in galaxies. And the Eddington limit that flattens the relation at the top is an active frontier: the winds and mass loss of very massive stars — which decide whether they end as neutron stars, black holes, or pair-instability explosions — all turn on how close they sit to their luminosity ceiling.

The physical picture

Picture a star as a leaky ball of hot gas held up against its own gravity by pressure. To support a larger mass the interior must be hotter and more compressed, because a deeper potential well demands a higher central pressure, and pressure in an ideal gas means temperature. A hotter, denser interior in turn lets radiation seep out faster — the escaping power is just the rate at which heat diffuses out of a self-gravitating sphere of a given mass. Run that logic through the structure equations and the luminosity turns out to depend on mass with astonishing steepness, and — the surprise — hardly at all on the star's radius or how vigorously it burns fuel.

Figure

A star as a leaky sphere of hot gas: gravity compresses the mass M so the core reaches the temperature T ∝ μM∕R needed for support, and that heat diffuses outward through the opacity κ as the luminosity L. Hydrostatic equilibrium plus radiative transport force L ∝ μ⁴M³∕κ, independent of the nuclear burning rate

The nuclear reactions play a subtler role than intuition suggests. Because rates depend on temperature like $T^{4}$ for the pp chain and up to $T^{18}$ for CNO, the core behaves as a near-perfect thermostat: it needs only a whisper of a temperature change to supply whatever power the structure demands. So fusion does not set the luminosity — it sets the size and surface temperature at which the star supplies it in thermal balance. The plot below is what the empiricist and the theorist agree on — the relation itself, across seven decades of luminosity.

Figure

The mass–luminosity relation on log–log axes: main-sequence stars follow L ∝ M³·⁵ across roughly seven decades of luminosity. Anchors are the Sun at (1, 1), Sirius A near (2.06, 25), a 10 M⊙ B star at (10, 3160), and a 0.2 M⊙ red dwarf near (0.2, 0.006); a 40 M⊙ O star falls below the extrapolated line as it approaches the Eddington limit

The core equations

The relation itself is a power law — a statement about slope on a log–log plot, with the exponent a running quantity, steepest near a solar mass and shallowest at the top.

L \propto M^{\alpha}, \qquad \alpha \approx 3.5 \qquad (1)

where: $L$ — bolometric luminosity, the total radiated power; $M$ — stellar mass; $\alpha$ — power-law index, $\approx 3\text{–}4$ across most of the main sequence. A factor of two in mass is roughly ten in light. For everyday work the relation is normalised to the Sun — the form used to weigh a star or estimate a lifetime.

\frac{L}{L_\odot} \approx \left(\frac{M}{M_\odot}\right)^{3.5} \qquad (2)

where: $L_\odot = 3.828\times10^{26}\,\mathrm{W}$ — solar luminosity; $M_\odot = 1.989\times10^{30}\,\mathrm{kg}$ — solar mass. A $2\,M_\odot$ star is predicted at $2^{3.5}\approx 11\,L_\odot$ (nearer $16$ with the local $\alpha\approx4$ ); a $10\,M_\odot$ star at $10^{3.5}\approx 3160\,L_\odot$ . Behind the power law stand the structure equations. The first is hydrostatic equilibrium: at every radius the pressure gradient holds up the weight of the overlying gas.

\frac{\mathrm{d}P}{\mathrm{d}r} = -\frac{G\,m(r)\,\rho}{r^{2}} \qquad (3)

where: $P$ — pressure at radius $r$ ; $m(r)$ — mass enclosed within $r$ ; $\rho$ — local density; $G$ — gravitational constant. More mass means a steeper gradient and a higher central pressure. The second is radiative diffusion, the resistance term: a large opacity $\kappa$ throttles the leakage and forces a steeper temperature drop to push the same luminosity through.

\frac{\mathrm{d}T}{\mathrm{d}r} = -\frac{3}{4ac}\,\frac{\kappa\rho}{T^{3}}\,\frac{L(r)}{4\pi r^{2}} \qquad (4)

where: $T$ — temperature; $\kappa$ — Rosseland mean opacity per unit mass; $a = 7.566\times10^{-16}\,\mathrm{J\,m^{-3}\,K^{-4}}$ — radiation constant; $c$ — speed of light; $L(r)$ — luminosity through $r$ . The third closes the system: an equation of state. For a main-sequence interior the ideal-gas law suffices, with the mean molecular weight $\mu$ counting how many particles each nucleon's worth of matter provides.

P = \frac{\rho\,k_B\,T}{\mu\,m_H} \qquad (5)

where: $k_B$ — Boltzmann constant; $\mu$ — mean molecular weight in units of $m_H$ ; $m_H$ — hydrogen atomic mass. As hydrogen fuses to helium, $\mu$ rises from about $0.6$ toward $1.3$ , and this is where that change enters $L$ . Feeding the three structure equations through the homology assumption yields the central analytic result — and its striking feature is what is absent: neither the radius nor the nuclear energy-generation rate appears.

L \propto \frac{\mu^{4}\,M^{3}}{\kappa} \qquad (6)

where: $\mu$ — mean molecular weight; its fourth power is why a star brightens as it ages at fixed mass; $\kappa$ — opacity, held constant here. Kramers opacity ( $\kappa\propto\rho\,T^{-3.5}$ ) adds mass dependence and steepens the slope toward $\alpha\approx4$ . At the top of the sequence the gas-pressure picture fails: radiation pressure dominates and electron scattering sets the resistance, so the luminosity runs into a ceiling — the Eddington luminosity, at which the outward radiation force balances gravity.

L_{\rm Edd} = \frac{4\pi G M c}{\kappa_{\rm es}} = \frac{8\pi G M m_H c}{\sigma_T (1+X)} \approx 3.8\times10^{4}\,\frac{M}{M_\odot}\,L_\odot \qquad (7)

where: $\kappa_{\rm es} = \sigma_T(1+X)/(2m_H) = 0.02(1+X)\,\mathrm{m^2\,kg^{-1}}$ — electron-scattering opacity, $0.034$ for $X=0.7$ ; $\sigma_T = 6.652\times10^{-29}\,\mathrm{m^2}$ — Thomson cross section; $X$ — hydrogen mass fraction. For pure ionised hydrogen ( $X=1$ , $\kappa=0.04$ ) this is the textbook $3.2\times10^{4}$ . Since $L_{\rm Edd}\propto M$ , the relation flattens toward $\alpha\to1$ here. Combining the fuel budget with the MLR gives the main-sequence lifetime, which falls steeply with mass because burn rate outruns fuel supply.

t_{\rm MS} \propto \frac{M}{L} \propto M^{\,1-\alpha} \approx M^{-2.5}, \qquad t_{\rm MS} \approx 10\,\mathrm{Gyr}\left(\frac{M}{M_\odot}\right)^{-2.5} \qquad (8)

where: $t_{\rm MS}$ — time fusing hydrogen in the core; $10\,\mathrm{Gyr}$ — the Sun's calibrated main-sequence lifetime. A $10\,M_\odot$ star lasts $\sim30\,\mathrm{Myr}$ ; a $0.5\,M_\odot$ dwarf, $\sim60\,\mathrm{Gyr}$ — longer than the Universe's present age.

Derivation AThe homology mass–luminosity scaling

The goal is to extract how luminosity scales with mass straight from the equations of stellar structure — hydrostatic equilibrium, the ideal-gas law, and radiative diffusion — without ever solving them, by assuming stars of different mass are homologous: structurally identical bodies differing only by an overall scale. Replacing every derivative by a ratio of characteristic central values turns the coupled equations into algebra, and the payoff is a surprise — both the stellar radius and the nuclear energy-generation rate cancel out, leaving luminosity fixed by gravity and photon transport alone.

Step 1. In hydrostatic equilibrium, replace $\mathrm{d}P/\mathrm{d}r \to P/R$ and $\rho \to M/R^{3}$ in equation (3): $P/R \sim GM(M/R^{3})/R^{2}$ , hence $P \sim GM^{2}/R^{4}$ . Why: this is the central pressure gravity demands to hold up mass $M$ in radius $R$ ; everything follows from paying that bill.

Step 2. Convert pressure to temperature with the ideal-gas law (5): $T \sim (\mu m_H/k_B)(P/\rho) \sim (\mu m_H G/k_B)(M/R)$ , so $T \propto \mu M/R$ . Why: temperature is what the gas trades for support, and it depends only on the well depth per radius $M/R$ times the mean molecular weight.

Step 3. Rearrange the diffusion equation (4) for the total luminosity, using $\mathrm{d}T/\mathrm{d}r \to T/R$ , $\rho \to M/R^{3}$ : $L \sim ac\,T^{4}R^{4}/(\kappa M)$ . Why: luminosity is the heat diffusing out per unit time, set by how steep the temperature is, how transparent the gas is, and how much matter the photons fight through.

Step 4. Substitute $T \propto \mu M/R$ from Step 2 into Step 3: $L \propto (ac/\kappa M)(\mu M/R)^{4}R^{4} = \mu^{4}M^{3}/\kappa$ . The two factors of $R^{4}$ cancel, so the radius vanishes — and so does the fusion rate. Why: the cancellation is the message — luminosity is not about a star's size or how hard it burns, but about how gravity heats it and how photons escape.

Step 5. Read off the exponent by regime: constant electron-scattering opacity gives $\alpha\approx3$ ; Kramers opacity $\kappa\propto\rho T^{-3.5}$ adds mass dependence and steepens it toward $\alpha\approx4$ for solar-type stars. Why: $\alpha\approx3.5$ is a mass-averaged compromise between these opacity laws — which is why it is quoted as an average, not a law. *Verified: reproduces the Carroll & Ostlie homology result $L\propto\mu^4 M^3/\kappa$ and is consistent with Kippenhahn, Weigert & Weiss Ch. 20; the $R^4$ cancellation and the $\alpha=3$ constant-opacity limit both recomputed.*

Go deeper · Derivation B · 1 minThe main-sequence lifetime from the fuel budget

Derivation BThe main-sequence lifetime from the fuel budget

The goal is to turn the mass–luminosity relation into a mass–lifetime relation by pitting two budgets against each other: the fuel a star carries, which grows only linearly with mass, and the rate at which it burns that fuel, its luminosity, which grows as a steep power of mass. Because a main-sequence star processes a roughly fixed fraction of its hydrogen at a roughly constant luminosity, lifetime is available energy divided by power; inserting the steep MLR makes it a sharply falling function of mass, and the Sun fixes the constant.

Step 1. Only a fraction $f\approx0.1$ of the star's mass — the hot core — is processed on the main sequence, and hydrogen-to-helium fusion releases a fraction $\epsilon\approx0.007$ of the rest mass, so $E_{\rm nuc} = f\,\epsilon\,M c^{2}$ . Why: the $0.7\%$ mass defect and the $\sim10\%$ core-mass fraction are the two numbers that turn a mass into an energy reservoir.

Step 2. With luminosity roughly constant over the phase, lifetime is reservoir over power: $t_{\rm MS} = E_{\rm nuc}/L = f\,\epsilon\,M c^{2}/L$ . Why: the thermostat holds $L$ nearly steady while fuel drains, so "tank size divided by burn rate" is a genuinely good approximation.

Step 3. Insert the MLR, $L\propto M^{\alpha}$ : $t_{\rm MS} \propto M/M^{\alpha} = M^{1-\alpha}$ , which for $\alpha\approx3.5$ is $t_{\rm MS}\propto M^{-2.5}$ . Why: linear growth of fuel is overwhelmed by the steep growth of luminosity — massive stars hoard more fuel but squander it far faster.

Step 4. Calibrate on the Sun: $t_\odot = (0.1)(0.007)(1.989\times10^{30})(2.998\times10^{8})^{2}/(3.828\times10^{26})\,\mathrm{s} \approx 3.3\times10^{17}\,\mathrm{s} \approx 10\,\mathrm{Gyr}$ . Why: real numbers turn the proportionality into an absolute clock — consistent with the Sun's known $4.6\,\mathrm{Gyr}$ age at roughly half its main-sequence life.

Step 5. Assemble the law: $t_{\rm MS}\approx 10\,\mathrm{Gyr}\,(M/M_\odot)^{-2.5}$ . A $10\,M_\odot$ star lives $\sim30\,\mathrm{Myr}$ ; a $0.5\,M_\odot$ dwarf, $\sim60\,\mathrm{Gyr}$ . Why: this is why star-forming regions breed supernovae within a few million years while the faintest dwarfs are effectively immortal. *Verified: solar lifetime recomputed as $3.3\times10^{17}\,\mathrm{s} = 10.4\,\mathrm{Gyr}$ from $f\epsilon Mc^2/L_\odot$ ; the $M^{-2.5}$ scaling and the $10\,M_\odot \to 32\,\mathrm{Myr}$ value match Carroll & Ostlie Ch. 10–11 and Prialnik Ch. 6.*

Go deeper · Derivation C · 1 minThe high-mass flattening toward the Eddington limit

Derivation CThe high-mass flattening toward the Eddington limit

The goal is to explain why the mass–luminosity relation flattens for the most massive stars, bending away from $L\propto M^{3.5}$ toward a nearly linear $L\propto M$ . Above a few tens of solar masses, radiation pressure supplies most of the support and electron scattering — an opacity independent of temperature and density — dominates the photon resistance. A ceiling luminosity then appears, the Eddington luminosity, at which the outward radiation force just balances gravity; a star pressing against it cannot radiate harder without unbinding its outer layers, so its luminosity is chained to its mass.

Step 1. In a very massive star gas pressure is a bit player; support comes from radiation, $P\approx P_{\rm rad}=\tfrac13 aT^{4}$ . Why: once radiation carries the pressure, the ideal-gas link between $T$ and support that steepened the homology relation is broken.

Step 2. Radiation pushes on matter through opacity: the radiative force per unit volume $(\kappa\rho/c)(L/4\pi r^{2})$ must not exceed gravity's pull $GM\rho/r^{2}$ . Why: if the push wins, the outer layers are driven off — the star cannot be more luminous than the balance point.

Step 3. Set the two equal; density and radius cancel, leaving $L_{\rm Edd} = 4\pi G M c/\kappa_{\rm es}$ . With constant electron-scattering opacity this is exactly linear in mass, $L_{\rm Edd}\propto M$ . Why: because $\kappa_{\rm es}$ carries no mass dependence, nothing steepens the relation — the ceiling itself rises only as fast as $M$ .

Step 4. Real massive stars are pushed toward the ceiling as $M$ grows, so the effective exponent slides from $\alpha\approx3.5$ through $3$ toward $1$ , and radiation-driven winds become a defining feature. Why: this is why the O-star end of the main sequence bows below the extrapolated power law and why such stars shed mass copiously. *Verified: $L_{\rm Edd}=4\pi GMc/\kappa_{\rm es}=8\pi GMm_Hc/[\sigma_T(1+X)]$ recomputed to $3.8\times10^4\,(M/M_\odot)\,L_\odot$ for $X=0.7$ ( $\kappa_{\rm es}=0.034\,\mathrm{m^2\,kg^{-1}}$ ); the pure-hydrogen value $3.2\times10^4$ recovered for $X=1$ .*

Limiting cases

The exponent in $L\propto M^{\alpha}$ is set by which physics dominates the opacity and pressure, and it changes across the sequence.

Case	Result	Meaning
Low mass, $M \lesssim 0.4\,M_\odot$	$L \propto M^{2.3}$	Fully convective M dwarfs; the radiative estimate breaks down, the slope is shallower, and degeneracy begins to matter.
Solar-type, $0.4 \lesssim M/M_\odot \lesssim 2$	$L \propto M^{4}$	Radiative envelopes with Kramers opacity give the steepest slope; the Sun anchors this segment.
Intermediate/high, $2 \lesssim M/M_\odot \lesssim 55$	$L \propto M^{3.5}$	Electron scattering takes over; the classic mass-averaged exponent.
Very massive, $M \gtrsim 55\,M_\odot$	$L \propto M^{1}$	Radiation-pressure supported; luminosity approaches the Eddington ceiling, linear in mass.

Common trapTreating α = 3.5 as a universal law that applies to any star

The single exponent $\alpha=3.5$ is a mass-averaged fit, not a law, and the MLR applies only to hydrogen-burning main-sequence stars. Giants, white dwarfs, brown dwarfs, and pre-main-sequence stars all deviate strongly — a red giant is enormously over-luminous for its mass, a white dwarf enormously under. Even on the main sequence the true slope runs from about $4$ near a solar mass down toward $1$ for the most massive stars. Two failure modes follow: using $3.5$ where the local slope is really $4$ (mis-weighing a $2\,M_\odot$ star), and extrapolating off the main sequence entirely (catastrophic for evolved stars). Always check the star is on the main sequence, then use the exponent for its mass band.

Worked examples

1 · A 10 M⊙ B star — luminosity, lifetime, and Eddington ratio (fully worked)

A main-sequence star has mass $M = 10\,M_\odot$ . Estimate its luminosity, its main-sequence lifetime, and how close it runs to the Eddington limit.

Step 1. Luminosity from the MLR (2): $L/L_\odot = 10^{3.5} = 1000\times3.162$ , so $L \approx \mathbf{3160\,L_\odot}$ — three thousandfold the Sun on a mass only ten times larger, the steepness of the relation in one number.

Step 2. Lifetime from (8): $t_{\rm MS} = 10\,\mathrm{Gyr}\times10^{-2.5} = 0.0316\,\mathrm{Gyr} \approx \mathbf{32\,Myr}$ — about $300$ times shorter a life. A galaxy's O and B stars are thus a running census of the last few tens of millions of years of star formation.

Step 3. Eddington ratio. From (7), $L_{\rm Edd} = 3.8\times10^{5}\,L_\odot$ , so $L/L_{\rm Edd} = 3160/(3.8\times10^{5}) = \mathbf{8.3\times10^{-3}}$ , under one percent. A $10\,M_\odot$ star is nowhere near the radiation-pressure ceiling — which is why it still obeys the steep $\alpha\approx3.5$ slope. Only above several tens of solar masses does $L/L_{\rm Edd}$ climb toward a few tenths and the relation bend over.

2 · Weighing a star from its light, and testing the relation

A detached eclipsing binary yields, from Kepler's third law, a primary mass $M = 2.0\,M_\odot$ ; photometry plus parallax give $L = 16\,L_\odot$ . Is this consistent with the MLR, and how would you invert it to weigh a single star?

Step 1. Near $2\,M_\odot$ the local slope is the steep solar-type value $\alpha\approx4$ , not the global $3.5$ . The prediction $L = 2^{4}\,L_\odot = \mathbf{16\,L_\odot}$ is an exact match, whereas the global $\alpha=3.5$ gives $2^{3.5}=11.3\,L_\odot$ , low by a third — the pitfall of using the mass-averaged exponent inside the steep solar-type band.

Step 2. Invert to weigh a single main-sequence star from its luminosity alone: $M/M_\odot = (L/L_\odot)^{1/4} = 16^{1/4} = \mathbf{2.0}$ . No companion needed — this photometric mass estimate underpins spectral-type mass calibrations and the initial mass function.

Step 3. The caveat that makes or breaks the method: it assumes the star is on the main sequence. Apply the same inversion to a red giant of $16\,L_\odot$ and you infer $2\,M_\odot$ when the true mass might be $1\,M_\odot$ — the MLR does not hold off the sequence.

3 · Dating a globular cluster from its turnoff

The main-sequence turnoff of a globular cluster sits at $M_{\rm to} = 0.9\,M_\odot$ — stars any more massive have already left the main sequence. Estimate the cluster's age.

Step 1. Coeval stars leave the main sequence most massive first, so the turnoff mass marks the star just now exhausting its core hydrogen, and its lifetime equals the cluster age: $t = 10\,\mathrm{Gyr}\times(0.9)^{-2.5}$ .

Step 2. Evaluate: $(0.9)^{-2.5} = 1.30$ , so $t \approx \mathbf{13\,Gyr}$ — close to the age of the Universe ( $13.8\,\mathrm{Gyr}$ ), and how globular clusters furnish an independent lower bound on cosmic age. Note the sensitivity: a turnoff at $1.0\,M_\odot$ would give $10\,\mathrm{Gyr}$ , so the turnoff mass must be pinned precisely, which is why real dating uses full isochrones rather than this single-power-law estimate.

Physics of the portrait

The mass–luminosity relation is where the empirical astronomy of binary stars met the new physics of radiative equilibrium, and one figure stands at that junction — the man who derived the relation from theory and discovered the luminosity ceiling that bends it at the top.

Origins · 1 minArthur Eddington and how this was discovered

Arthur Eddington

1882–1944 · the theory of radiative equilibrium and the internal constitution of the stars

Arthur Stanley Eddington was the architect of the modern theory of stellar interiors. Through the 1910s and 1920s he established that a star is a ball of gas held up by gas and radiation pressure, with energy carried outward by the diffusion of photons — the picture from which the mass–luminosity relation follows. His 1924 paper "On the relation between the masses and luminosities of the stars" derived, from radiative equilibrium alone, that luminosity should climb steeply with mass, and — to his surprise — that even giant stars fell on the same relation, showing ordinary gas physics governed them all. Along the way he identified the luminosity at which radiation pressure overwhelms gravity, the Eddington limit that still bears his name, and his 1926 monograph The Internal Constitution of the Stars gathered it into the foundational text of the field. He worked before the source of stellar energy was known, arguing against the physics of his day that the interiors must be hot enough for subatomic energy release, and lived to see Bethe's reactions vindicate him — the rare theorist whose 1919 eclipse expedition also confirmed general relativity.

History

1911

— Halm notes that more massive stars are systematically more luminous

1919

— Hertzsprung and Russell sharpen the empirical relation from binary-star masses

1924

— Eddington derives the mass–luminosity relation from radiative equilibrium

1926

— Eddington's The Internal Constitution of the Stars codifies stellar structure and the luminosity limit

1939

— Bethe details the proton–proton chain and CNO cycle that power the main sequence

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., Ch. 7 (binary systems and stellar masses) and Ch. 10–11 (stellar structure and the main sequence) · Kippenhahn, Weigert & Weiss, Stellar Structure and Evolution, 2nd ed., Ch. 20–22 (homology and the MLR) · Hansen, Kawaler & Trimble, Stellar Interiors, 2nd ed. · Prialnik, An Introduction to the Theory of Stellar Structure and Evolution, Ch. 6 · Binney & Tremaine, Galactic Dynamics, 2nd ed. (mass-to-light ratios).

Provenance. The structure equations and the homology argument follow Carroll & Ostlie Ch. 10–11, Kippenhahn/Weigert/Weiss Ch. 20–22, and Prialnik Ch. 6. A verifier should recompute: (1) the $R^{4}$ cancellation giving $L\propto\mu^4 M^3/\kappa$ ; (2) the solar lifetime $\approx3.3\times10^{17}\,\mathrm{s}\approx10\,\mathrm{Gyr}$ from $f\epsilon Mc^2/L_\odot$ with $f=0.1$ , $\epsilon=0.007$ ; (3) $L_{\rm Edd}=4\pi GMc/\kappa_{\rm es}=8\pi GMm_Hc/[\sigma_T(1+X)]\approx3.8\times10^{4}(M/M_\odot)L_\odot$ for $X=0.7$ (and $3.2\times10^{4}$ for $X=1$ ); (4) $10^{3.5}\approx3160$ , $2^{4}=16$ , $16^{1/4}=2$ , $0.9^{-2.5}\approx1.30$ . Constants are CODATA/IAU values. Note: the microscopic Eddington form is $8\pi GMm_Hc/[\sigma_T(1+X)]$ , not $4\pi$ — the opacity carries a factor $1/(2m_H)$ .

Prerequisites. The Main Sequence · The Hertzsprung–Russell Diagram · The Proton–Proton Chain · The CNO Cycle · Gamow Peak and Reaction Rates. Leads to. Post-Main-Sequence Evolution and the Red Giant Branch · Helium Burning and the Triple-Alpha Process · Advanced Burning Stages in Massive Stars · The Chandrasekhar Mass.

Ask the tutor: Want to see the homology cancellation play out, or weigh a specific binary or date a real cluster from its turnoff? Ask and we will take it step by step.

Why is a star's luminosity set by gravity and photon diffusion rather than by how fast it burns fuel?

Because the core acts as a thermostat. Nuclear rates depend on temperature so steeply — $T^{4}$ for the pp chain, up to $T^{18}$ for CNO — that the star needs only a minute temperature adjustment to supply whatever power leaks out. The structure equations fix that leakage: given a mass, gravity sets the interior temperature and opacity sets how fast photons escape, and together they demand a definite luminosity that fusion self-regulates to match. Boost the reaction rate by hand and the star merely expands and cools until diffusion again carries away the demanded power. Luminosity is a structural property; fusion sets the radius and surface temperature at which the star delivers it.

Why isn't the exponent α a single universal constant?

Because different physics governs the opacity and the pressure in different mass ranges. For solar-type stars, Kramers opacity ( $\kappa\propto\rho T^{-3.5}$ ) steepens the slope toward $\alpha\approx4$ ; for intermediate and high masses, constant electron-scattering opacity settles it near $3.5$ ; for the most massive stars the luminosity approaches the Eddington ceiling and flattens toward $1$ ; and below $0.4\,M_\odot$ the fully convective, partly degenerate dwarfs give a shallower, uncertain slope near $2.3$ . The quoted $\alpha=3.5$ is a mass-averaged compromise — fine for rough work, wrong by enough to matter inside any one band.

How were stellar masses actually measured to calibrate the relation?

Almost entirely from binary stars, because a mass can only be measured through its gravity. Kepler's third law relates a binary's period and separation to the total mass, and the ratio of the two stars' orbital speeds splits that total into individual masses. Detached eclipsing binaries are the gold standard: the eclipses give radii and inclination, the radial-velocity curves give the masses, all from geometry, to a percent or better. Luminosities then come from photometry plus a distance — now delivered in bulk by Gaia. Every point on the empirical MLR is, at bottom, a star whose gravity was weighed against a companion.

Why do massive stars live such short lives when they carry so much more fuel?

Because burn rate outruns fuel supply, and badly. The accessible fuel scales linearly with mass, $E\propto M$ , but the luminosity scales as $L\propto M^{3.5}$ , so lifetime $t\propto M/L\propto M^{-2.5}$ falls steeply. A $10\,M_\odot$ star carries ten times the Sun's fuel but burns it more than three thousand times faster, emptying its tank about three hundred times sooner — thirty million years versus ten billion. Extra fuel buys a spectacularly brighter, spectacularly briefer star.

Does the mass–luminosity relation apply to giants, white dwarfs, or brown dwarfs?

No — it is strictly a main-sequence relation. A red giant burns hydrogen in a shell around an inert core and is hugely over-luminous for its mass; a white dwarf has no fusion at all and sits far below. Brown dwarfs never sustain hydrogen fusion and cool continuously, and pre-main-sequence stars are still contracting. The relation holds only in the stable, core-hydrogen-burning configuration that defines the main sequence — exactly the regime the derivation assumes. Off it, weighing a star from its light gives nonsense.

What is the Eddington luminosity, and does any real star reach it?

The Eddington luminosity is the brightness at which radiation's outward force on matter, exerted through opacity, exactly balances gravity — radiate harder and the outer layers are pushed off. For electron scattering it is $L_{\rm Edd}=4\pi GMc/\kappa_{\rm es}\approx3.8\times10^{4}(M/M_\odot)\,L_\odot$ , linear in mass. Ordinary stars sit well below it — the Sun at one part in ten thousand, a $10\,M_\odot$ star under a percent. But the most massive stars run at a substantial fraction of Eddington, which is why they drive powerful winds, shed mass throughout their lives, and why the MLR flattens toward $L\propto M$ at the top.