Tully-Fisher Relation

The Tully-Fisher relation is the tightest scaling law spiral galaxies obey, and it wears two faces. Operationally it is a distance indicator: measure how fast a disk spins — a quantity that does not care how far away it is — read the galaxy's true luminosity off a calibrated line, compare with how bright it looks, and out drops the distance. Structurally it is a statement about galaxy formation: the light a spiral emits is fixed, to a few tenths of a magnitude, by the depth of the gravitational well it sits in, and that well is dug mostly by invisible dark matter. Tully and Fisher published it in 1977 as a correlation between optical luminosity and 21-cm line width; forty years of work has sharpened it into one of the sturdiest facts in extragalactic astronomy. Two refinements matter. First, the relation steepens and tightens from blue light to the near-infrared, because red light traces the old stars that carry most of the mass while blue light is polluted by patchy young populations and dust. Second — and more deeply — replacing luminosity by total baryonic mass, stars plus cold gas, straightens the relation into a clean $M_{\rm bar}\propto V^{4}$ power law that runs unbroken across five decades in mass, down into the gas-rich dwarfs where the luminosity version fails. That baryonic form is where the physics is cleanest, and it is the sharpest test astronomers have of how baryons settle inside dark halos — and of whether dark matter is even the right explanation.

Key takeaways

The Tully-Fisher relation ties a spiral's luminosity to its rotation speed, $L\propto V_{\max}^{\alpha}$ with $\alpha\approx3$ – $4$ ; because $V_{\max}$ is distance-independent it turns rotation into an absolute magnitude and hence a distance.
The rotation speed is read from the width of the 21-cm HI line, corrected for tilt: $W_{\rm obs}=2V_{\max}\sin i$ , so a nearly face-on galaxy shows almost no width however fast it spins.
The slope steepens and the scatter shrinks from blue to near-infrared, and the relation is cleanest of all when luminosity is replaced by baryonic mass — the Baryonic Tully-Fisher Relation, $M_{\rm bar}\propto V_{\max}^{4}$ .
A simple virial argument — circular velocity measures enclosed mass, plus a constant surface brightness and mass-to-light ratio — forces the fourth power, tying a galaxy's light to its dark-matter well.
Calibrated by Cepheids, the TFR reaches clusters out to $\sim100$ Mpc and feeds measurements of the Hubble constant.
Key idea: for a rotationally supported disk in equilibrium the circular velocity measures the enclosed mass, $V^2\sim GM/R$ ; folding in a roughly constant disk surface brightness and mass-to-light ratio eliminates the size and forces $L\propto V^4$ , chaining a galaxy's visible light to the depth of its dark-matter-dominated potential well.

The core insight

For a rotationally supported disk in dynamical equilibrium the circular velocity measures the enclosed mass, $V^2\sim GM/R$ ; combined with an approximately constant disk surface brightness and mass-to-light ratio this forces $L\propto V^4$ , tying a galaxy's light to the depth of its (dark-matter-dominated) potential well.

Where this lives today

The Tully-Fisher relation is a working instrument, not a museum piece. On the distance ladder it is calibrated locally by Cepheid variables and reaches spiral galaxies far past where individual stars can be resolved, mapping the local velocity field — peculiar velocities and bulk flows — out to clusters at $\sim100$ Mpc and feeding the contested value of the Hubble constant. Wide-field HI surveys such as ALFALFA and the SKA pathfinders measure line widths for tens of thousands of galaxies, while integral-field surveys (SAMI, MaNGA) replace the single global width with resolved velocity fields and push the relation to higher redshift. Its baryonic form is the headline galaxy-scale test in the dark-matter-versus-MOND debate: the SPARC database of $\sim175$ disk galaxies shows the relation holding with scatter so small it strains hierarchical galaxy-formation models to reproduce without fine-tuning, yet is a parameter-free prediction of modified dynamics. It also remains the template for the Faber-Jackson relation and the fundamental plane of ellipticals, which span the whole Hubble sequence with a common virial logic.

The physical picture

Picture a spiral disk as a spinning platter held together by gravity. Out at large radius the rotation speed levels off to a constant $V_{\max}$ — the flat rotation curve — and that value is a direct readout of the total mass enclosed, most of it dark: a deeper well spins faster. Now add two empirical regularities: spiral disks share roughly the same central surface brightness (Freeman's law), and, in the near-infrared, roughly the same mass-to-light ratio. Together these say a deeper well holds proportionally more baryons, hence more light. The Tully-Fisher relation is the quantitative form of that chain — and it is far tighter than any galaxy-to-galaxy comparison has a right to be, the clue that baryons and dark matter grow up tightly coupled. The rotation speed is not measured directly; it is inferred from the 21-cm line of neutral hydrogen. Gas on the approaching edge is blueshifted, gas on the receding edge redshifted, and the integrated line spans the full spread of projected velocities — a double-horned profile of width $2V_{\max}\sin i$ . The tilt $i$ enters because only the line-of-sight component of the rotation Doppler-shifts the gas, as the first figure shows.

Figure

A tilted spiral disk turns rotation into a double-horned line: gas on the approaching edge is blueshifted and the receding edge redshifted, so for i = 60° and V_max = 200 km/s the 21-cm profile spans the full projected width W_obs = 2 V_max sin i = 346 km/s

Why the fourth power, and not the bare virial second power? A single $V^2\propto M/R$ argument gives mass, but turning mass into light means eliminating the size $R$ . Constant surface brightness does exactly that — it makes $R\propto L^{1/2}$ — and the two powers of $R$ hidden in that substitution promote $V^2$ into $V^4$ . Bluer bands scatter more because they are dominated by transient, dust-reddened young stars; near-infrared light tracks the old stellar mass, which is what gravity weighs, and adding the cold gas removes the last systematic. The second figure is that punchline: spirals and dwarfs on one line of slope four.

Figure

The baryonic Tully-Fisher relation: stellar plus gas mass versus flat rotation speed for spirals and gas-rich dwarfs lands on one power law M_bar = 50 V_max⁴ (slope 4) across five decades in mass, while the stars-only relation (dashed amber) bends downward where dwarfs become gas-dominated and light stops tracking mass

The core equations

The relation is stated most practically as a straight line in the magnitude versus log-line-width plane, the form a calibrator actually fits. The slope $a$ is negative — brighter (more negative $M$ ) galaxies have larger widths — and the zero point $b$ is nailed down by galaxies whose distances are known independently, chiefly through Cepheid variables.

M = a\,\log_{10}\!\big(2\,V_{\max}\big) + b \qquad (1)

where: $M$ — absolute magnitude in a chosen passband; $V_{\max}$ — maximum, inclination-corrected rotation velocity; $2V_{\max}$ — the intrinsic full line width $W$ (approaching plus receding); $a$ — slope (negative; roughly $-6$ to $-10$ depending on band); $b$ — zero point set by primary calibrators. A wider line means a more luminous galaxy, and the same fit run in a redder band gives a steeper, tighter line. Converting magnitudes to luminosity turns the straight line into a power law. The two forms are the same statement, with $\alpha=-0.4\,a$ because a magnitude is $-2.5\log_{10}L$ .

L \propto V_{\max}^{\,\alpha}, \qquad \alpha \approx 3\text{–}4 \qquad (2)

where: $L$ — total luminosity in the band; $\alpha$ — power-law slope, $\alpha=-0.4\,a$ . The fourth power is the deep case (near-infrared and baryonic); the shallower slope $\sim3$ is the dust-and-youth-contaminated blue. The velocity that enters must be the intrinsic rotation speed, not the projected width you observe. Because only the line-of-sight component of the rotation Doppler-shifts the gas, the observed width shrinks with the disk's tilt and must be divided back out.

V_{\max} = \frac{W_{\rm obs}}{2\sin i} \qquad (3)

where: $W_{\rm obs}$ — observed 21-cm line width in velocity units; $i$ — disk inclination ( $i=90°$ edge-on shows the full rotation, $i=0°$ face-on shows none). Inclination comes from the apparent axis ratio, and small- $i$ galaxies make poor targets because the correction blows up. The physics that makes the relation exist is the circular-velocity law: for a mass in equilibrium the rotation speed at radius $R$ measures everything enclosed inside it, dark matter included.

V_c^{2}(R) = \frac{G\,M(<R)}{R} \qquad (4)

where: $V_c$ — circular velocity at radius $R$ ; $M(<R)$ — total enclosed mass; $G$ — Newton's constant. On the flat part of the curve $V_c\to V_{\max}$ is constant, so it weighs the halo, not just the visible disk. To turn mass into light we need the disk's light profile. Spiral disks are exponential, and integrating that profile gives the total luminosity in terms of a central brightness and a scale length.

L = 2\pi\,I_0\,R_d^{2} \qquad (5)

where: $I_0$ — central surface brightness (nearly constant among spirals — Freeman's law, $\mu_0\approx21.65$ B-mag arcsec $^{-2}$ ); $R_d$ — disk scale length. Constant $I_0$ makes $L\propto R_d^2$ , so size and luminosity are not independent — the fact that eliminates $R$ from the virial law. For the baryonic version, luminosity is converted to stellar mass through a mass-to-light ratio and the cold gas is added on, scaling the measured neutral-hydrogen mass up for helium.

M_{\rm bar} = M_\star + M_{\rm gas} = \Upsilon_\star\,L + 1.33\,M_{\rm HI} \qquad (6)

where: $\Upsilon_\star\equiv M_\star/L$ — stellar mass-to-light ratio (steadiest in the near-infrared); $M_{\rm HI}$ — neutral-hydrogen mass from the 21-cm flux; $1.33$ — factor folding in primordial helium. In bright spirals $M_\star$ dominates and $L$ suffices; in dwarfs the gas term takes over, which is why light alone fails there. Written as a mass relation, the baryonic form is a clean quartic with a nearly universal normalization — the single most important number in this subject.

M_{\rm bar} = A\,V_{\max}^{4}, \qquad A \approx 50\ M_\odot\,\mathrm{km^{-4}\,s^{4}} \qquad (7)

where: $A$ — normalization (McGaugh et al. 2000; observed values $\sim47$ – $50$ ); $M_{\rm bar}$ — total baryonic mass in solar masses; $V_{\max}$ — flat rotation velocity in km s $^{-1}$ . The tiny scatter and constant $A$ across five decades make this a precision benchmark for galaxy formation and for MOND. Finally, the relation earns its keep as a distance tool through the distance modulus: the calibrated $M$ minus the observed apparent magnitude gives the distance directly.

\mu = m - M = 5\log_{10}\!\left(\frac{d}{10\,\mathrm{pc}}\right) \qquad (8)

where: $\mu$ — distance modulus; $m$ — apparent magnitude, corrected for extinction; $M$ — absolute magnitude from equation (1); $d$ — luminosity distance. $V_{\max}$ delivers $M$ without knowing $d$ , so subtracting the observed $m$ inverts straight to distance.

Derivation ADeriving the fourth-power law from a constant surface brightness

The goal is to show where the exponent 4 comes from, using only the circular-velocity law and two empirical regularities of spiral disks. We want luminosity as a function of rotation velocity alone, so both the mass and the size must be eliminated: the virial law, a constant mass-to-light ratio, and Freeman's constant surface brightness supply the three relations that close the system. Watch how the single power of size hidden in the surface-brightness law doubles the exponent from two to four.

Step 1. Read the enclosed mass off the rotation curve: at a characteristic radius $R\sim R_d$ the circular-velocity law gives $V_{\max}^{2}=GM/R$ , so $M=V_{\max}^{2}R/G$ , where $M$ is the total dark-plus-baryonic mass inside $R$ . Why: this is the only place dynamics enters — the flat rotation speed scales the depth of the potential well.

Step 2. Tie mass to light with a roughly constant effective mass-to-light ratio, $M=\Upsilon\,L$ . Why: this assumes galaxies convert well depth into light in a regular way — true if halo-to-baryon ratios and stellar populations are similar, which is what makes the relation tight.

Step 3. Use Freeman's law: spirals share a central surface brightness $I_0$ , so $L=2\pi I_0 R^2$ and therefore $R=(L/2\pi I_0)^{1/2}$ . Why: this eliminates size in favour of luminosity — for real disks size and light are not independent, and this hidden $R\propto L^{1/2}$ is what upgrades the exponent.

Step 4. Substitute $M=\Upsilon L$ and $R\propto L^{1/2}$ into the mass equation from Step 1:

\Upsilon L = \frac{V_{\max}^{2}}{G}\left(\frac{L}{2\pi I_0}\right)^{1/2}.

Why: everything is now in $L$ and $V_{\max}$ only; the rest is collecting powers of $L$ .

Step 5. Gather the $L$ terms — $L/L^{1/2}=L^{1/2}\propto V_{\max}^2$ — and square:

L^{1/2}\propto \frac{V_{\max}^{2}}{\Upsilon}\,(2\pi I_0)^{1/2} \quad\Longrightarrow\quad L\propto \frac{V_{\max}^{4}}{\Upsilon^{2}\,I_0}.

Why: squaring the $V^2$ relation is where the fourth power is born; the zero point is set by $\Upsilon$ and $I_0$ , so observed slopes of 3–4 track the band-dependent drift of $\Upsilon$ , and the relation's tightness measures how nearly constant $\Upsilon$ and $I_0$ really are.

Go deeper · Derivation B · 1 minThe baryonic relation from a flat rotation curve

Derivation BThe baryonic relation from a flat rotation curve

The goal is a cleaner route to $M_{\rm bar}\propto V^{4}$ that never assumes constant surface brightness — only a characteristic mean baryonic surface density and a flat rotation curve — since it is the baryonic form that holds across the widest range of galaxies. We evaluate the circular-velocity law at the edge of the baryon distribution, replace the dynamical mass by the baryonic mass through a universal baryon fraction, and eliminate the radius with a fixed surface density, so the fourth power again appears from squaring.

Step 1. On the flat part of the curve the rotation speed is constant; evaluate the circular-velocity law at the outer edge $R$ of the baryons: $V_{\max}^{2}=GM_{\rm dyn}(<R)/R$ . Why: the flat curve means one number $V_{\max}$ characterizes the whole well, so the choice of $R$ affects only the normalization.

Step 2. Relate baryonic to dynamical mass with a roughly universal baryon fraction $\mathcal{F}$ : $M_{\rm bar}=\mathcal{F}\,M_{\rm dyn}$ , so $M_{\rm dyn}=M_{\rm bar}/\mathcal{F}$ . Why: this assumes every disk collects a similar share of its halo's mass as baryons — the content behind the small scatter.

Step 3. Introduce a characteristic mean baryonic surface density $\Sigma\equiv M_{\rm bar}/(\pi R^2)$ , giving $R=(M_{\rm bar}/\pi\Sigma)^{1/2}$ . Why: this replaces the unknown size by the mass itself, using a disk property that varies far less than mass does.

Step 4. Substitute Steps 2 and 3 into Step 1:

V_{\max}^{2}=\frac{G}{\mathcal{F}}\,M_{\rm bar}\left(\frac{\pi\Sigma}{M_{\rm bar}}\right)^{1/2}=\frac{G}{\mathcal{F}}\,(\pi\Sigma)^{1/2}\,M_{\rm bar}^{1/2}.

Why: the radius is gone, leaving $V_{\max}^2$ linked to the square root of the baryonic mass — the relation that squares into a quartic.

Step 5. Square both sides to isolate the mass:

M_{\rm bar}=\frac{V_{\max}^{4}}{G^{2}\,\pi\Sigma\,\mathcal{F}^{-2}}\;\propto\;V_{\max}^{4}.

Why: with $\Sigma$ and $\mathcal{F}$ fixed the normalization is constant, giving $M_{\rm bar}=A\,V_{\max}^4$ with $A\approx50\,M_\odot\,{\rm km^{-4}\,s^4}$ — and because nothing here referred to starlight, the relation survives into gas-rich dwarfs.

Go deeper · Derivation C · 1 minTurning a line width into a distance

Derivation CTurning a line width into a distance

The goal is the full recipe that converts a radio line width and a photograph into a distance in megaparsecs — the operation that puts the Tully-Fisher relation on the extragalactic distance ladder. Given a calibration fixed by galaxies with Cepheid distances, the rotation speed predicts the absolute magnitude, the apparent magnitude is cleaned of extinction, and their difference is the distance modulus — keeping the distance-dependent and distance-independent quantities apart until the last line.

Step 1. Measure the intrinsic rotation speed by correcting the observed 21-cm width for tilt: $V_{\max}=W_{\rm obs}/(2\sin i)$ , with small further corrections for turbulent and instrumental broadening. Why: this is the one measurement that carries no distance information — what lets the method bootstrap a distance at all.

Step 2. Predict the absolute magnitude from the calibrated relation in the same band as the photometry: $M=a\,\log_{10}(2V_{\max})+b$ . Why: the calibration transfers the fourth-power physics into the galaxy's true luminosity without ever knowing its distance.

Step 3. Correct the apparent magnitude for foreground and internal dust and for the K-correction: $m_0=m_{\rm obs}-A_\lambda-K(z)$ . Why: extinction dims a galaxy exactly as distance does, so an uncorrected $m$ masquerades as extra distance — the dominant systematic in practice.

Step 4. Form the distance modulus and invert:

\mu=m_0-M \quad\Longrightarrow\quad d=10^{\,(\mu+5)/5}\ \mathrm{pc}.

Why: only now do the distance-dependent $m_0$ and distance-independent $M$ meet; their difference is pure distance, good to $\sim15$ – $20\%$ per galaxy and far better averaged over a cluster, reaching $\sim100$ Mpc and feeding $H_0$ .

Limiting cases

How the relation behaves across passbands and mass regimes:

Case	Result	Meaning
Optical (B band)	$L_B \propto V_{\max}^{\,\sim 3}$	Shallower slope, larger scatter ( $\sim0.4$ – $0.5$ mag): dust and young blue stars vary galaxy to galaxy.
Near-infrared (H, K, 3.6 μm)	$L_H \propto V_{\max}^{\,\sim 4}$	Steeper and tighter ( $\sim0.3$ mag): traces the old stellar mass, nearly dust-free.
Baryonic (stars + gas)	$M_{\rm bar} \propto V_{\max}^{\,4}$	The cleanest power law; unbroken from giant spirals to dwarfs, smallest scatter.
Low-mass / dwarf regime	$L$ bends; $M_{\rm bar}\propto V^4$ holds	Dwarfs are gas-dominated, so light underweights mass and the luminosity relation curls away.
MOND / low-acceleration limit	$M_{\rm bar}=V_{\max}^4/(G\,a_0)$	Modified dynamics predicts the quartic and its normalization, $a_0\approx1.2\times10^{-10}\ \mathrm{m\,s^{-2}}$ .

Common trapA narrow 21-cm line does not mean a slowly rotating galaxy

The velocity that enters the Tully-Fisher relation is the intrinsic maximum rotation speed, never the raw observed line-of-sight width. Because $W_{\rm obs}=2V_{\max}\sin i$ , a nearly face-on disk ( $i\to0$ ) shows an almost vanishing line width no matter how fast it truly spins — the rotation is projected away along the line of sight. You must divide by $\sin i$ , with $i$ from the apparent axis ratio, and also remove turbulent, instrumental, and redshift broadening. Neglecting the $\sin i$ correction is the single most common source of catastrophic error, and it is why nearly face-on galaxies, where the correction is large and uncertain, are deliberately excluded from samples.

Worked examples

1 · The distance to a spiral galaxy (fully worked)

A spiral has a 21-cm line width $W_{\rm obs}=346$ km s $^{-1}$ , inclination $i=60°$ , and extinction-corrected I-band magnitude $m_I=12.0$ . Using the I-band calibration $M_I=-21.0-7.68\,(\log_{10}W^i-2.5)$ , where $W^i=2V_{\max}$ is the inclination-corrected width, find its distance.

Step 1. Correct the width for tilt: $V_{\max}=W_{\rm obs}/(2\sin i)=346/(2\sin60°)=346/1.732=\mathbf{200\ km\,s^{-1}}$ , so the intrinsic width is $W^i=2V_{\max}=400$ km s $^{-1}$ .

Step 2. Predict the absolute magnitude. With $\log_{10}400=2.602$ , $M_I=-21.0-7.68\,(2.602-2.5)=\mathbf{-21.78}$ — intrinsically luminous, as its fast rotation demands.

Step 3. Distance modulus: $\mu=m_I-M_I=12.0-(-21.78)=\mathbf{33.78}$ .

Step 4. Invert to distance: $d=10^{(\mu+5)/5}=10^{(38.78)/5}=10^{7.76}=5.7\times10^{7}$ pc. So d ≈ 57 Mpc — well beyond the reach of resolved stars, from a radio line width and one photograph.

2 · Why a gas-rich dwarf needs the baryonic relation (fully worked)

A dwarf spiral rotates at $V_{\max}=50$ km s $^{-1}$ , with a stellar mass $M_\star=6\times10^{7}\,M_\odot$ and neutral-hydrogen mass $M_{\rm HI}=1.9\times10^{8}\,M_\odot$ . Show that its stars alone fall far below the Tully-Fisher line but its baryons land on it.

Step 1. Assemble the cold-gas mass including helium: $M_{\rm gas}=1.33\,M_{\rm HI}=1.33\times1.9\times10^{8}=\mathbf{2.53\times10^{8}\,M_\odot}$ — outweighing the stars four to one, the hallmark of a dwarf.

Step 2. Total baryonic mass: $M_{\rm bar}=M_\star+M_{\rm gas}=6\times10^{7}+2.53\times10^{8}=\mathbf{3.13\times10^{8}\,M_\odot}$ .

Step 3. Compare with the prediction $M_{\rm bar}=50\,V_{\max}^4=50\times50^{4}=\mathbf{3.13\times10^{8}\,M_\odot}$ — a dead-on match.

Step 4. Now try stars only: $M_\star=6\times10^{7}\,M_\odot$ is a factor of 5 below the line. A luminosity-based relation effectively sees only $M_\star$ and would place this galaxy far off the line — which is exactly why the optical TFR bends downward for dwarfs while the baryonic version stays straight. Baryons obey the quartic; light alone does not.

3 · The acceleration scale hidden in the normalization (fully worked)

MOND predicts the baryonic relation as $M_{\rm bar}=V_{\max}^4/(G\,a_0)$ with a universal acceleration $a_0$ . Show that the standard value $a_0=1.2\times10^{-10}$ m s $^{-2}$ reproduces the observed normalization.

Step 1. The predicted normalization is $A=1/(G\,a_0)$ . Insert SI values: $G\,a_0=(6.674\times10^{-11})(1.2\times10^{-10})=8.0\times10^{-21}$ , so $A=\mathbf{1.25\times10^{20}}$ kg s $^4$ m $^{-4}$ .

Step 2. Convert to galaxy units, using $1\,(\mathrm{km\,s^{-1}})^4=10^{12}\,(\mathrm{m\,s^{-1}})^4$ and $1\,M_\odot=1.989\times10^{30}$ kg: $A=1.25\times10^{20}\times10^{12}/1.989\times10^{30}=\mathbf{63\,M_\odot\,(km\,s^{-1})^{-4}}$ .

Step 3. The measured normalization is $A\approx47$ – $50\,M_\odot\,(\mathrm{km\,s^{-1}})^{-4}$ , so the parameter-free MOND prediction sits within a factor $\sim1.3$ . That agreement is why the baryonic Tully-Fisher relation is the sharpest galaxy-scale test between dark matter and modified dynamics.

Physics of the portrait

The Tully-Fisher relation is where measurement outran theory — a number that worked for years before anyone could say why, and whose deepest form still divides interpretations.

Origins · 1 minTully-Fisher relation and how this was discovered

Tully-Fisher relation

the L ∝ V⁴ scaling of rotating disks, R. B. Tully and J. R. Fisher, 1977 · extragalactic distance scale and galaxy dynamics

The relation carries two names, but its roots run deeper than 1977: as early as 1922 Ernst Öpik estimated Andromeda's distance from its rotation and mean density — the first hint that how fast a galaxy spins encodes how much it weighs and how far it lies. The modern relation was announced by R. Brent Tully and J. Richard Fisher in 1977 (A&A 54, 661), correlating the optical luminosities of spirals with their 21-cm line widths as a distance indicator, a year after Sandra Faber and Robert Jackson found the pressure-supported analogue for ellipticals, $L\propto\sigma^4$ . The physical scaffolding was already in place: Ken Freeman had shown in 1970 that spiral disks share a nearly universal central surface brightness, the ingredient that turns the bare virial $V^2$ into the observed $V^4$ , while Vera Rubin's flat rotation curves revealed that the velocity setting the relation is a readout of the dark halo, not the visible disk. The decisive modern refinement came in 2000, when Stacy McGaugh and collaborators showed that folding cold gas into a total baryonic mass straightens the relation into a clean $M_{\rm bar}\propto V^4$ running unbroken across five decades in mass, including the gas-rich dwarfs where luminosity fails. That baryonic form, with its tiny scatter and no-free-parameter agreement with the MOND acceleration scale, is now the battleground on which dark matter and modified dynamics are tested — a demand on any theory of galaxy formation.

History

1922

— Öpik infers Andromeda's distance from its rotation and mean density

1970

— Freeman finds spiral disks share a nearly constant central surface brightness

1976

— Faber and Jackson report L ∝ σ⁴ for elliptical galaxies

1977

— Tully and Fisher publish the luminosity–line-width relation for spirals

2000

— McGaugh and collaborators establish the baryonic Tully-Fisher relation

Going further

Texts. Carroll & Ostlie, An Introduction to Modern Astrophysics, 2nd ed., ch. 24–25 (spiral galaxies, distance indicators) · Binney & Tremaine, Galactic Dynamics, 2nd ed., ch. 1 & 4 (scaling relations, disk dynamics) · Mo, van den Bosch & White, Galaxy Formation and Evolution, ch. 11 (disk structure and scaling laws) · Ryden, Introduction to Cosmology, ch. 6 (the distance ladder) · Dodelson, Modern Cosmology (secondary distance indicators and $H_0$ ). Key results used throughout: slopes $\alpha\approx3$ – $4$ and scatter $\sim0.3$ – $0.4$ mag; the exponential-disk integral $L=2\pi I_0 R_d^2$ ; the $\sin i$ tilt geometry; the algebra giving $L\propto V^4$ ; the worked distance ( $V_{\max}=200$ km s $^{-1}$ , $M_I=-21.78$ , $d\approx57$ Mpc); the dwarf baryonic assembly ( $M_{\rm bar}=3.13\times10^8\,M_\odot$ matching $50V^4$ ); and the MOND normalization ( $A\approx63\,M_\odot\,{\rm km^{-4}\,s^4}$ from $a_0=1.2\times10^{-10}$ m s $^{-2}$ , compared with the observed $\approx47$ – $50$ ).

Prerequisites. Galactic Rotation Curves · Neutral Hydrogen 21-cm Line and Galactic Kinematics · Spiral Galaxies and Density Wave Theory · Evidence for Dark Matter in Galaxies. Leads to. Galactic Dark Matter Halos and Density Profiles · Elliptical Galaxies and the de Vaucouleurs Profile · Hubble Sequence and Galaxy Morphology.

Ask the tutor: Ask me why constant surface brightness turns the virial $V^2$ into $V^4$ , or to work a distance or baryonic-mass example with your own numbers.

Why is the exponent 4 specifically, rather than 2 or 3?

The bare virial law gives $V^2\propto M/R$ , a second power, but that is a relation for mass, not light. Converting to luminosity means eliminating the size $R$ , and spiral disks obey a size-luminosity relation ( $L\propto R^2$ at constant surface brightness) that lets you write $R\propto L^{1/2}$ . Substituting folds an extra factor of $V^2$ into the algebra, so $L^{1/2}\propto V^2$ and hence $L\propto V^4$ — "virial squared, with size scaled out." Real slopes drift between 3 and 4 because the mass-to-light ratio and surface brightness are only approximately constant, by an amount that depends on the band.

Is the Tully-Fisher relation evidence for dark matter or for MOND?

Both camps claim it. In the dark-matter picture the rotation speed reflects the halo, and the tightness reflects a nearly universal baryon fraction and a regular way baryons settle inside halos — but reproducing the small scatter without fine-tuning is a genuine challenge for simulations. In MOND the relation is no coincidence: modified dynamics predicts $M_{\rm bar}=V^4/(Ga_0)$ with a fixed acceleration $a_0$ , matching both exponent and normalization with no free parameters. The relation does not settle the debate; it sharpens it, because any theory must explain not just that it exists but why its scatter is so small.

How is the disk inclination actually measured, and why does it matter so much?

Inclination is estimated from the apparent shape: a circular disk seen at angle $i$ projects to an ellipse of axis ratio $b/a=\cos i$ (with a small correction for finite disk thickness). Since $W_{\rm obs}=2V_{\max}\sin i$ , recovering the true speed means dividing by $\sin i$ . The catch is that for nearly face-on galaxies $\sin i$ is small and highly sensitive to errors in $b/a$ , so the correction amplifies any mistake — a 5° error near face-on can swamp the measurement. This is why Tully-Fisher samples restrict themselves to reasonably inclined disks, typically $i\gtrsim45°$ .

What ultimately limits the accuracy of a Tully-Fisher distance?

Several effects, roughly in order. Internal dust extinction is the worst: it dims a galaxy exactly as distance does and depends on inclination and morphology, so it must be modelled carefully. The inclination correction adds error, especially for less-inclined disks. Intrinsic scatter — about 0.3 mag even in the near-infrared — sets a floor of roughly 15% per galaxy, and the whole ladder inherits its Cepheid zero point. Averaging over cluster members beats down the random error, which is why cluster distances are far more reliable and why the method is trusted out to $\sim100$ Mpc.

Why is the baryonic version considered more fundamental than the luminosity version?

Because it comes closest to a pure statement about mass and dynamics, with the least astrophysical contamination. Luminosity is a proxy for stellar mass, and a poor one where gas dominates or where young stars and dust distort the light; the baryonic mass $M_\star+1.33M_{\rm HI}$ counts the actual matter that gravitates. Empirically the payoff is dramatic: the luminosity relation bends and scatters for low-mass galaxies, while the baryonic relation stays a single straight power law of slope 4 across five decades in mass with the smallest scatter of any galaxy scaling law — exactly the form a fundamental relation should take, reaching from giant spirals to the faintest dwarfs.

How does the Tully-Fisher relation relate to Faber-Jackson and the fundamental plane?

They are the same virial idea told for different galaxies. Spirals are rotationally supported, so ordered rotation $V$ sets the potential and gives $L\propto V^4$ (Tully-Fisher). Ellipticals are pressure-supported, so the velocity dispersion $\sigma$ plays that role, giving $L\propto\sigma^4$ (Faber-Jackson). The fundamental plane refines the elliptical case by adding size and surface brightness as a second axis, tightening it just as surface brightness tightens the spiral case. All three share a common origin in $V^2\sim GM/R$ plus structural regularities, which is why a single logic spans the whole Hubble sequence.