The Two-Body Problem and Reduced Mass

Strictly speaking, the Moon does not orbit the Earth. Both bodies orbit their common center of mass — a point about 1,700 km beneath the Earth's surface — so our planet executes a small monthly circle while the Moon executes a large one, always on opposite sides of the pivot. The Sun does the same dance with Jupiter, wobbling by roughly its own radius; astronomers read exactly this reflex wobble in distant stars to discover planets they cannot see. The mathematics behind the dance looks forbidding at first: two particles, six coordinates, coupled by a mutual force. This article shows that one change of variables dissolves the whole difficulty. Split the motion into the center of mass $\mathbf{R}$ and the relative coordinate $\mathbf{r} = \mathbf{r}_1 - \mathbf{r}_2$ , and the problem falls apart into two independent pieces: a center of mass that drifts in a straight line, carrying no information about the interaction, and a single fictitious particle of reduced mass $\mu = m_1m_2/(m_1+m_2)$ moving in the potential $V(r)$ about a fixed center. Two bodies become one; six coordinates become three. The payoff is everything that follows in this track. All of central-force theory, the effective potential, and the Kepler problem are built on the one-body problem this reduction delivers — every result holding for real two-body systems under the single substitution $m \to \mu$ . That substitution weighs binary stars, corrects Kepler's third law, and shifted hydrogen's spectral lines by just enough to reveal deuterium. One coordinate change, a century of physics.

Key takeaways

Two bodies interacting through a potential $V(|\mathbf{r}_1-\mathbf{r}_2|)$ that depends only on their separation have a Lagrangian that separates exactly into a center-of-mass (CM) part and a relative-coordinate part.
The CM moves like a free particle of total mass $M = m_1 + m_2$ : $\ddot{\mathbf{R}} = 0$ . Its motion is trivial and is removed by working in the CM frame.
The relative motion is an equivalent one-body problem: a fictitious particle of reduced mass $\mu = m_1m_2/M$ obeying $\mu\ddot{\mathbf{r}} = -\nabla V(r)$ about a fixed center.
The individual positions follow from $\mathbf{r}_1 = \mathbf{R} + (m_2/M)\,\mathbf{r}$ and $\mathbf{r}_2 = \mathbf{R} - (m_1/M)\,\mathbf{r}$ : each body traces a scaled copy of the relative orbit about the CM.
The reduced mass satisfies $1/\mu = 1/m_1 + 1/m_2$ and $\mu \le \min(m_1, m_2)$ ; for $m_2 \gg m_1$ it tends to $m_1$ , recovering the fixed-center approximation, while equal masses give $\mu = m/2$ .
Applied to gravity, the reduction yields the exact third law $T^2 = 4\pi^2 a^3 / G(m_1+m_2)$ — the total mass, not just the central one — the basis for weighing binary stars.
Key idea: the transformation to CM and relative coordinates diagonalizes the kinetic energy, so the Lagrangian splits into a free CM particle plus a one-body problem of mass μ — two bodies become one.

The core insight

The transformation $(\mathbf{r}_1,\mathbf{r}_2)\to(\mathbf{R},\mathbf{r})$ diagonalizes the kinetic energy, so the Lagrangian becomes $L=\tfrac12 M\dot{\mathbf{R}}^2+\tfrac12\mu\dot{\mathbf{r}}^2-V(r)$ — a free CM particle plus a one-body problem in the reduced mass $\mu$ .

Where this lives today

Exoplanet hunting is the two-body reduction industrialized. A star with an unseen planet orbits the pair's CM at $a_\star = (m_p/M)a$ — typically meters per second of reflex velocity — and high-resolution spectrographs like ESPRESSO read that wobble in the Doppler shift of stellar lines to better than 30 cm/s, hunting Earth-mass planets whose reflex signal is under 10 cm/s. Gaia's astrometry attacks the same wobble positionally, and its 2026 data release is expected to yield thousands of binaries and giant planets weighed by exactly Worked Example 2's logic: period plus orbit sizes in, masses out. The same separation now underwrites gravitational-wave astronomy. A black-hole binary is a two-body problem whose relative coordinate spirals inward as the system radiates; the waveform detected by LIGO–Virgo–KAGRA is controlled at leading order by the chirp mass $\mathcal{M} = \mu^{3/5}M^{2/5}$ — a specific cocktail of this article's two mass parameters. Every one of the hundreds of merger events now catalogued had its masses extracted by matching templates parametrized in $(\mu, M)$ : the reduced-mass reduction, holding up under conditions Newton never imagined, one waveform at a time.

The physical picture

Imagine two children of unequal weight on a spinning seesaw, pivoting about its balance point. The heavier child sits close to the pivot and traces a small circle; the lighter child sits far out and traces a big one. Both complete their circles in the same time, always diametrically opposite. Every gravitational pair — Earth and Moon, two stars in a binary, the Sun and Jupiter — is that seesaw. The pivot is the center of mass, and the sizes of the two circles are in inverse proportion to the masses.

Figure

The two-body dance — both bodies orbit the common center of mass, tracing scaled copies of one relative orbit

Now think of the relative coordinate $\mathbf{r}$ as a rigid stick joining the two bodies. The interaction cares only about the length and orientation of this stick — never about where its midpoint sits in space. So the stick's dynamics unfold as if one particle danced in the potential $V(r)$ , while the midpoint (more precisely, the mass-weighted balance point) drifts off uniformly and irrelevantly. That is the entire content of the reduction, and the reduced mass $\mu$ is simply the effective inertia the stick presents: how hard it is to change the separation, given that both ends recoil when either is pushed. The formula $\mu = m_1m_2/(m_1+m_2)$ interpolates between two familiar regimes. When the masses are equal, $\mu = m/2$ — pushing the two ends apart is twice as easy as pushing one body against a wall, because both ends give way. When one mass dominates, $\mu \to m_{\text{light}}$ and the heavy body becomes an effectively immovable force center — which is why treating the Sun as nailed down works so well for planets. Energy bookkeeping makes the split vivid: the total kinetic energy is exactly the CM kinetic energy plus the relative-motion kinetic energy, with no cross term. That vanishing cross term is the whole magic, and it happens precisely because $\mathbf{R}$ and $\mathbf{r}$ carry the one mass-weighting that diagonalizes the kinetic energy.

Figure

The coordinate change — trade two position vectors for the center of mass R and the separation r

The separation, step by step

Start from the honest description. Two particles, positions $\mathbf{r}_1$ and $\mathbf{r}_2$ , interacting through a potential that depends only on how far apart they are — gravity between star and planet, Coulomb attraction between proton and electron, a spring between two gliders. The Lagrangian is kinetic minus potential:

L = \tfrac{1}{2}m_1\dot{\mathbf{r}}_1^2 + \tfrac{1}{2}m_2\dot{\mathbf{r}}_2^2 - V(|\mathbf{r}_1 - \mathbf{r}_2|) \qquad (1)

where: $V$ — the interaction potential, a function of the separation $|\mathbf{r}_1-\mathbf{r}_2|$ alone. As written, the two coordinates are coupled through $V$ : neither Euler–Lagrange equation can be solved without the other. Six coupled degrees of freedom. The move that untangles them is to stop tracking the bodies individually and track instead the balance point and the gap:

\mathbf{R} = \frac{m_1\mathbf{r}_1 + m_2\mathbf{r}_2}{M}, \qquad \mathbf{r} = \mathbf{r}_1 - \mathbf{r}_2, \qquad M = m_1 + m_2 \qquad (2)

where: $\mathbf{R}$ — the center of mass, weighted toward the heavier body, and $\mathbf{r}$ — the relative coordinate pointing from body 2 to body 1. These definitions invert cleanly (substitute $\mathbf{r}_2 = \mathbf{r}_1 - \mathbf{r}$ into the CM definition and solve):

\mathbf{r}_1 = \mathbf{R} + \frac{m_2}{M}\,\mathbf{r}, \qquad \mathbf{r}_2 = \mathbf{R} - \frac{m_1}{M}\,\mathbf{r} \qquad (3)

where: the mass fractions $m_2/M$ and $m_1/M$ set how far each body sits from the CM — each body's lever arm is proportional to its partner's mass, which is just the seesaw balance condition. Equation (3) is also the recipe for reconstructing the real trajectories once the relative problem is solved, so it earns its keep twice. The last definition packages the effective inertia of the gap:

\mu = \frac{m_1 m_2}{m_1 + m_2}, \qquad \frac{1}{\mu} = \frac{1}{m_1} + \frac{1}{m_2} \qquad (4)

where: $\mu$ — the reduced mass. The reciprocal form is the one to memorize: reduced masses combine like resistors in parallel, which makes the limits transparent — the smaller mass dominates the combination, and $\mu$ is always smaller than either. With the cast assembled, the central theorem:

Derivation AExact separation of the two-body Lagrangian

We are proving that the coordinate change (2) splits the Lagrangian (1) into two completely independent pieces — a free particle of mass $M$ at $\mathbf{R}$ , plus a particle of mass $\mu$ in the potential $V(r)$ — with no approximation whatsoever. The strategy is direct substitution: differentiate the inverse transformation (3), insert into the kinetic energy, and expand the squares. The entire theorem lives or dies on one event: the cross terms $\dot{\mathbf{R}}\cdot\dot{\mathbf{r}}$ must cancel exactly, and they do so only because $\mathbf{R}$ carries precisely the mass-weighting $m_1\mathbf{r}_1 + m_2\mathbf{r}_2$ — any other weighting (the geometric midpoint, say) leaves a coupling behind. Watch also where the reduced mass is born: it appears uninvited, as the coefficient the algebra assigns to $\dot{\mathbf{r}}^2$ .

Step 1. Differentiate the inverse transformation (3):

\dot{\mathbf{r}}_1 = \dot{\mathbf{R}} + \frac{m_2}{M}\dot{\mathbf{r}}, \qquad \dot{\mathbf{r}}_2 = \dot{\mathbf{R}} - \frac{m_1}{M}\dot{\mathbf{r}}

Step 2. Insert into the kinetic energy and expand both squares:

T = \tfrac{1}{2}m_1\Big(\dot{\mathbf{R}} + \tfrac{m_2}{M}\dot{\mathbf{r}}\Big)^2 + \tfrac{1}{2}m_2\Big(\dot{\mathbf{R}} - \tfrac{m_1}{M}\dot{\mathbf{r}}\Big)^2

Step 3. Collect the three kinds of term. The $\dot{\mathbf{R}}^2$ terms give $\tfrac12(m_1+m_2)\dot{\mathbf{R}}^2 = \tfrac12 M\dot{\mathbf{R}}^2$ . The cross terms give $\big(m_1\tfrac{m_2}{M} - m_2\tfrac{m_1}{M}\big)\dot{\mathbf{R}}\cdot\dot{\mathbf{r}} = 0$ . The $\dot{\mathbf{r}}^2$ terms give $\tfrac12\big(\tfrac{m_1m_2^2 + m_2m_1^2}{M^2}\big)\dot{\mathbf{r}}^2 = \tfrac12\tfrac{m_1m_2}{M}\dot{\mathbf{r}}^2$ :

T = \tfrac{1}{2}M\dot{\mathbf{R}}^2 + \tfrac{1}{2}\mu\,\dot{\mathbf{r}}^2

Why the cross terms die: the two coefficients are $+m_1m_2/M$ and $-m_1m_2/M$ — equal and opposite by construction of the mass-weighted CM. This is the only step where the definition of $\mathbf{R}$ matters, and it is exactly what "mass-weighted" is for. Note the coefficient of $\dot{\mathbf{r}}^2$ : $m_1m_2/M = \mu$ . The reduced mass is not postulated — the algebra manufactures it.

Step 4. Add the potential. Since $V$ depends only on $|\mathbf{r}|$ , the full Lagrangian splits into an $\mathbf{R}$ -sum and an $\mathbf{r}$ -sum with no shared variable:

L = \underbrace{\tfrac{1}{2}M\dot{\mathbf{R}}^2}_{\text{CM: free}} + \underbrace{\Big(\tfrac{1}{2}\mu\dot{\mathbf{r}}^2 - V(r)\Big)}_{\text{relative: one body in } V} \qquad (5)

Step 5. Read off the equations of motion. $\mathbf{R}$ is cyclic (it never appears undifferentiated), so its Euler–Lagrange equation gives $\mathrm{d}(M\dot{\mathbf{R}})/\mathrm{d}t = 0$ — total momentum conserved, CM drifting uniformly. The $\mathbf{r}$ equation gives Newton's law for one particle:

\ddot{\mathbf{R}} = 0, \qquad \mu\ddot{\mathbf{r}} = -\nabla V(r) \qquad (6)

Why this is the payoff: the second equation is exactly the equation of a single particle of mass $\mu$ in an external central potential with the force center nailed to the origin. Every theorem of one-body central-force mechanics now applies to the genuine two-body problem verbatim, under $m \to \mu$ . *Verified: matches Goldstein §3.1, Eqs. (3.4)–(3.5), and Landau–Lifshitz §13; the cross-term cancellation is checked by expanding with explicit coefficients in Step 3.*

The separation theorem has a quiet corollary worth displaying, because observers live in the CM frame. Setting $\mathbf{R} = 0$ in (3), the individual positions become fixed multiples of the relative coordinate:

\mathbf{r}_1(t) = \frac{m_2}{M}\,\mathbf{r}(t), \qquad \mathbf{r}_2(t) = -\frac{m_1}{M}\,\mathbf{r}(t) \qquad (7)

where: the minus sign keeps the two bodies always on opposite sides of the CM, collinear with it. Both real orbits are photocopies of the relative orbit — same shape, same period, scaled by the partner's mass fraction and flipped for body 2. If $\mathbf{r}(t)$ traces an ellipse of semi-major axis $a$ , body 1 traces one of $a_1 = (m_2/M)a$ and body 2 one of $a_2 = (m_1/M)a$ , with $a_1 + a_2 = a$ . Measuring the ratio of the two orbit sizes hands you the mass ratio $a_1/a_2 = m_2/m_1$ directly — the first half of how binary stars are weighed. The energy splits the same clean way,

E = \tfrac{1}{2}M\dot{\mathbf{R}}^2 + \tfrac{1}{2}\mu\dot{\mathbf{r}}^2 + V(r) \qquad (8)

where: the first term is a constant you can boost away, and everything dynamical lives in the relative terms — which is why "the energy of the orbit" always means the relative part. The total angular momentum splits identically into $M\,\mathbf{R}\times\dot{\mathbf{R}}$ plus the internal $\mu\,\mathbf{r}\times\dot{\mathbf{r}}$ , the latter conserved on its own in the CM frame and feeding directly into the central-force machinery.

What the reduction does to Kepler's third law

Here is the reduction earning real money. For gravity, $V(r) = -Gm_1m_2/r$ , and the equivalent one-body problem can be fed straight into the standard Kepler results — with one twist that textbooks using a fixed Sun never show you.

Go deeper · Derivation B · 1 minKepler's third law with the correct total mass

Derivation BKepler's third law with the correct total mass

The goal is the exact period–semi-major-axis relation for a genuine two-body gravitational system, $T^2 = 4\pi^2a^3/G(m_1+m_2)$ , and in particular to see why the total mass appears — not the central body's mass, as the naive fixed-Sun derivation suggests. The strategy: write the relative equation of motion from (6), divide through by $\mu$ , and watch the mass combination $Gm_1m_2/\mu$ collapse to $GM$ . We then import the standard Kepler-orbit period formula (derived in the Kepler problem) for a particle around a fixed center of effective strength $GM$ . The subtlety to watch: $\mu$ cancels out of the trajectory entirely, but it does so by promoting the effective central mass from $m_2$ to $M$ — the light body's own mass re-enters through the door of the heavy body's motion.

Step 1. Write the relative equation of motion for gravity:

\mu\ddot{\mathbf{r}} = -\frac{Gm_1m_2}{r^2}\,\hat{\mathbf{r}}

Step 2. Divide by $\mu$ and simplify the coefficient:

\ddot{\mathbf{r}} = -\frac{Gm_1m_2}{\mu\,r^2}\,\hat{\mathbf{r}} = -\frac{GM}{r^2}\,\hat{\mathbf{r}}

Why the collapse: $m_1m_2/\mu = m_1m_2\cdot\frac{m_1+m_2}{m_1m_2} = M$ . The relative coordinate accelerates as if orbiting a fixed body of the combined mass — because when body 1 falls toward body 2, body 2 is simultaneously falling toward body 1, and the gap closes at the sum of the two rates.

Step 3. Import the standard result for a particle about a fixed center of strength $GM$ : an elliptical orbit of semi-major axis $a$ has period $T = 2\pi\sqrt{a^3/GM}$ . Squaring:

T^2 = \frac{4\pi^2}{G(m_1+m_2)}\,a^3 \qquad (9)

where: $a$ — the semi-major axis of the relative orbit. This is Kepler's third law as nature actually enforces it: the constant of proportionality contains the total mass.

Step 4. Quantify the correction. For Earth around the Sun the factor is $(1 + m_\oplus/m_\odot) \approx 1.000003$ — the naive law errs by three parts per million, which is why Kepler never noticed. For a binary star with comparable masses the correction is order unity: ignoring it gets stellar masses wrong by factors of two. Solar-system convenience, not a law of nature. Verified: matches Morin Ch. 7 and Taylor Ch. 8; the m₁m₂/μ = M identity is checked by direct algebra, and the Earth–Sun correction 3×10⁻⁶ equals the mass ratio m⊕/m⊙ = 3.0×10⁻⁶.

Go deeper · Derivation C · 1 minReconstructing the individual trajectories

Derivation CReconstructing the individual trajectories

Solving the relative problem gives $\mathbf{r}(t)$ — but no real object sits at $\mathbf{r}$ . The goal is to recover the two physical trajectories and establish the geometric picture used throughout binary-star astronomy: two similar conics about the common CM, in lockstep. The strategy is almost embarrassingly short — work in the CM frame and apply the inverse transformation — but the geometric consequences (scale factors, shared focus, equal periods) are exactly what observers measure, so we spell them out honestly. Watch the sign in body 2's relation: it is what keeps the bodies on opposite sides of the CM at every instant.

Step 1. Go to the CM frame: $\mathbf{R} = 0$ for all time (the CM moves uniformly by (6), so this is an inertial frame — no fictitious forces needed).

Step 2. Apply the inverse transformation (3) with $\mathbf{R} = 0$ :

\mathbf{r}_1(t) = \frac{m_2}{M}\,\mathbf{r}(t), \qquad \mathbf{r}_2(t) = -\frac{m_1}{M}\,\mathbf{r}(t)

Why the geometry follows instantly: each physical position is a constant scalar multiple of $\mathbf{r}(t)$ . Scaling a conic by a constant gives a similar conic; the minus sign is a point reflection through the CM. So both bodies trace conics similar to the relative orbit.

Step 3. Read off the scales. If the relative orbit is an ellipse of semi-major axis $a$ and eccentricity $e$ , then body 1's orbit has $a_1 = (m_2/M)a$ , body 2's has $a_2 = (m_1/M)a$ , both with the same eccentricity $e$ , both with a focus at the CM, and $a_1 + a_2 = a$ . Why the heavier body's orbit is smaller: its lever arm scales with its partner's mass fraction — the seesaw again. Jupiter, at a thousandth of a solar mass, drags the Sun around an orbit a thousandth the size of its own.

Step 4. Confirm the synchrony. Both bodies' positions are fixed multiples of the same $\mathbf{r}(t)$ , so both sweep their orbits with identical period $T$ — always diametrically opposite the CM, never anywhere else. Verified: consistent with Goldstein §3.1 and the standard binary-star treatment; the scale factors sum to a and reproduce the a₁/a₂ = m₂/m₁ mass-ratio method.

Which equation answers which question

you want	use	notes
the effective one-body inertia	$\mu = m_1m_2/M$ , eq. (4)	parallel-resistor form $1/\mu = 1/m_1 + 1/m_2$
the equations of motion	$\ddot{\mathbf{R}} = 0$ , $\mu\ddot{\mathbf{r}} = -\nabla V$ , eq. (6)	exact, no approximation
real positions from the relative orbit	eq. (3) or CM-frame eq. (7)	scale by the partner's mass fraction
period of a gravitational binary	$T^2 = 4\pi^2a^3/GM$ , eq. (9)	$a$ = relative-orbit semi-major axis, $M$ = total
mass ratio of a binary	$a_1/a_2 = m_2/m_1$	from the two observed orbit sizes
when to bother with μ	error $\sim m_1/m_2$	3×10⁻⁶ for Earth–Sun, 1.2% for Moon–Earth

Limiting cases

Case	Result	Meaning
equal masses, $m_1 = m_2 = m$	$\mu = m/2$	positronium, equal binaries; each orbit is half the relative orbit
one mass dominant, $m_2 \gg m_1$	$\mu \to m_1$ , $\mathbf{R} \to \mathbf{r}_2$	heavy body ≈ fixed center; the textbook one-body picture recovered
general bound	$\mu \le \min(m_1, m_2)$	equality only for an infinitely heavy partner
symmetric form	$1/\mu = 1/m_1 + 1/m_2$	resistors in parallel; the lighter mass dominates
Earth–Sun	$\mu = m_\oplus(1 - 3.0\times10^{-6})$	fixed-Sun approximation good to 3 ppm
Moon–Earth	$\mu = 0.988\,m_{\text{Moon}}$	1.2% correction — already matters for lunar theory

Common trapThe orbiting body has its own mass, so use its mass in the orbit equations

This is the classic two-body error, and it is wrong twice over. First: the object that traces the relative orbit is the fictitious reduced particle, and its mass is $\mu$ — not $m_1$ , not $m_2$ . Using the Moon's full mass in the Earth–Moon relative dynamics is a 1.2% error; using an electron's full mass in hydrogen misplaces every spectral line by one part in 1836 (a huge error by spectroscopic standards — it is the difference between hydrogen and deuterium). Second, the companion confusion: the relative orbit $\mathbf{r}(t)$ is not the orbit of either real body. Each body traces a scaled-down copy about the CM — similar shape, same eccentricity and period, size set by the partner's mass fraction, eq. (7). The full-mass, fixed-center picture is a limit ( $m_2 \gg m_1$ ), not the theory. When the masses are comparable — binary stars, positronium, many exoplanet systems around light stars — only $\mu$ and the CM-frame geometry give right answers.

Worked examples

1 · The Earth–Moon seesaw (confidence builder)

Given $m_\oplus = 5.97\times10^{24}\ \mathrm{kg}$ , $m_{\text{Moon}} = 7.34\times10^{22}\ \mathrm{kg}$ , and mean separation $d = 3.84\times10^8\ \mathrm{m}$ : find the reduced mass of the system, and locate the center of mass. Earth's radius is $6.37\times10^6\ \mathrm{m}$ .

Step 1. Total mass: $M = 5.97\times10^{24} + 0.0734\times10^{24} = 6.045\times10^{24}\ \mathrm{kg}$ .

Step 2. Reduced mass from (4): $\mu = \dfrac{(5.97\times10^{24})(7.34\times10^{22})}{6.045\times10^{24}} = \boxed{7.25\times10^{22}\ \mathrm{kg}}$ . Compare with the Moon's mass: $\mu/m_{\text{Moon}} = 7.25/7.34 = 0.988$ — the reduced mass is 1.2% below the lighter body, exactly the fractional correction $\approx m_{\text{Moon}}/m_\oplus = 1/81$ .

Step 3. CM location from Earth's center, using the mass-fraction rule (3): $d_\oplus = (m_{\text{Moon}}/M)\,d = (7.34\times10^{22}/6.045\times10^{24}) \times 3.84\times10^8 = 0.01214 \times 3.84\times10^8 = \boxed{4.66\times10^6\ \mathrm{m} = 4{,}660\ \mathrm{km}}$ .

Step 4. Interpret: Earth's radius is 6,370 km, so the pivot of the Earth–Moon seesaw lies about 1,700 km beneath the Earth's surface. The Earth genuinely orbits this interior point once a month — an 9,300-km-diameter monthly wobble that lunar laser ranging measures routinely, and the reason "the Moon orbits the Earth" is only 98.8% true.

2 · Weighing a binary star (exam level)

A visual binary has an orbital period $T = 20\ \mathrm{yr}$ and a relative orbit with semi-major axis $a = 10\ \mathrm{AU}$ . The two stars' individual orbits about the CM have measured semi-major axes in the ratio $a_1/a_2 = 1.5$ . Find both stellar masses. (In solar units, Kepler's third law (9) reads $T^2[\mathrm{yr}] = a^3[\mathrm{AU}]/(M[M_\odot])$ .)

Step 1. Total mass from the exact third law (9) in solar units: $M = a^3/T^2 = 10^3/20^2 = 1000/400 = \boxed{2.5\ M_\odot}$ . Note this needed only the relative orbit and the period — and it delivers the total mass, not either star's.

Step 2. Mass ratio from the orbit-size ratio, eq. (7): $\dfrac{a_1}{a_2} = \dfrac{m_2}{m_1} = 1.5$ , so $m_2 = 1.5\,m_1$ . The star with the bigger orbit is the lighter one — the seesaw rule.

Step 3. Solve the pair: $m_1 + 1.5\,m_1 = 2.5\ M_\odot$ gives $m_1 = \boxed{1.0\ M_\odot}$ and $m_2 = \boxed{1.5\ M_\odot}$ .

Step 4. Cross-checks. Individual orbits: $a_1 = (m_2/M)a = (1.5/2.5)(10) = 6\ \mathrm{AU}$ , $a_2 = 4\ \mathrm{AU}$ ; ratio $6/4 = 1.5$ ✓ and sum $6 + 4 = 10\ \mathrm{AU} = a$ ✓. Reduced mass, for the record: $\mu = (1.0)(1.5)/2.5 = 0.6\ M_\odot$ . This two-step logic — total mass from period and relative size, ratio from the individual sizes — is the method by which stellar masses are known at all; every mass in an astronomy textbook traces back to some binary weighed exactly this way.

Origins · 1 minJoseph-Louis Lagrange and how this was discovered

Joseph-Louis Lagrange

1736–1813 · Turin, Berlin, Paris

Lagrange turned mechanics from geometry into analysis. Self-taught in Turin, he sent Euler results on the calculus of variations at nineteen; Euler delayed his own publication so the younger man could have the credit. His Mécanique analytique (1788) famously contains not a single diagram — the two-body separation performed in this article, with its cyclic CM coordinate and conserved momentum, is precisely his machinery of generalized coordinates at work. Tackling the three-body problem, where no such clean reduction exists, he won the 1772 Paris Academy prize for finding the special solutions now called Lagrange points — including L2, where the James Webb Space Telescope now sits, a million miles out along the Sun–Earth line, parked in the mathematics of a 250-year-old prize essay. Napoleon made him a count; he is buried in the Panthéon.

History

1687

— Newton's Principia solves the gravitational two-body problem, proving both bodies orbit their common center of gravity, which "either is at rest, or moves uniformly in a right line."

1772

— Lagrange's prize essay on the three-body problem finds the five equilibrium points of a rotating two-body system — the Lagrange points that now anchor space telescopes.

1889

— Poincaré's King Oscar prize memoir shows the three-body problem admits no general closed solution, discovering deterministic chaos in the gap between two bodies and three.

1932

— Urey announces deuterium, discovered from faint companion spectral lines shifted by the isotope's different reduced mass — the μ-correction made visible, and a Nobel Prize.

1995

— Mayor and Queloz detect 51 Pegasi b from its star's reflex wobble about the star–planet center of mass: the two-body reduction, run backwards, becomes the exoplanet Doppler method.

Going further

Texts. Morin, Introduction to Classical Mechanics, Ch. 7 (Central Forces) · Goldstein, Poole & Safko, Classical Mechanics, 3rd ed., §3.1 · Landau & Lifshitz, Mechanics, 3rd ed., §13 · Taylor, Classical Mechanics, Ch. 8.

Provenance. All equations are standard textbook results. A verifier should: (1) confirm the kinetic-energy cross-term cancellation algebraically; (2) check $1/\mu=1/m_1+1/m_2$ and the limits $\mu\to m/2$ (equal) and $\mu\to m_1$ ( $m_2\gg m_1$ ); (3) verify the exact Kepler third law $T^2=4\pi^2 a^3/[G(m_1+m_2)]$ by the substitution $Gm_1m_2/\mu=GM$ ; (4) confirm the individual-orbit scale factors $m_2/M$ , $m_1/M$ . Sources: Goldstein §3.1, Landau–Lifshitz §13, Morin Ch. 7, Taylor Ch. 8. The hydrogen reduced-mass shift $\sim 1/1836$ uses the standard proton-to-electron mass ratio $m_p/m_e\approx 1836$ . Worked-example numbers on this page ( $\mu = 7.25\times10^{22}$ kg and CM at 4,660 km for Earth–Moon; $M = 2.5\,M_\odot$ , $m_1 = 1.0\,M_\odot$ , $m_2 = 1.5\,M_\odot$ for the binary) are illustrative arithmetic recomputed from the boxed formulas.

Builds on: Lagrangian mechanics · Newton's laws and conservation of momentum · Vector kinematics. Leads to: Central force fundamentals · Angular momentum conservation in central fields · Reduction to the radial equation · The Kepler problem · Kepler's third law.

Is the reduced particle a real object? What is actually located at position r?

Nothing is located at $\mathbf{r}$ — it is the gap, not a place where matter sits. The reduced particle is a bookkeeping fiction: a single imaginary object whose position is the separation vector $\mathbf{r}_1 - \mathbf{r}_2$ and whose mass $\mu$ is chosen so that Newton's second law for it reproduces the true relative dynamics exactly. The fiction is harmless because the map back to reality, eq. (7), is trivial: each real body sits at a fixed multiple of $\mathbf{r}$ on either side of the CM. The situation is exactly like the center of mass itself, which is also generally a point where no matter sits (the CM of a donut is in the hole). Mechanics repeatedly earns simplicity by promoting useful combinations of coordinates to the status of "particles" — the whole subject of normal coordinates does the same thing for oscillators.

Why is the reduced mass always smaller than both masses? I would have guessed some kind of average.

Because it measures how easy it is to change the separation, and the separation changes when either body gives way. Push the two bodies apart with a given force: body 1 accelerates one way at $F/m_1$ , body 2 the other way at $F/m_2$ , and the gap grows at the sum of the rates, $F/m_1 + F/m_2 = F/\mu$ . More response per unit force means less effective inertia — so $\mu$ is smaller than either mass, not an average of them. The parallel-resistor form $1/\mu = 1/m_1 + 1/m_2$ encodes precisely this "two escape routes" logic, and it also shows the lighter body dominates: an infinitely heavy partner contributes nothing to $1/\mu$ (it never budges), leaving $\mu = m_{\text{light}}$ .

So does the Moon orbit the Earth or not? What is the honest statement?

The honest statement: the Earth and Moon each orbit their common center of mass, which lies about 4,660 km from Earth's center — inside the planet, about 1,700 km below the surface. Because the pivot is inside the Earth, "the Moon orbits the Earth" is a good cartoon: the Moon's orbit is 81 times larger than Earth's. But Earth's share of the dance is real and measured — the planet circles that interior point monthly, and lunar laser ranging (bouncing light off retroreflectors left on the Moon) tracks the geometry to millimeters. For the Sun–Jupiter pair the CM sits just outside the solar surface, so even the cartoon fails: the Sun visibly orbits a point above its own photosphere. The two-body reduction is indifferent to where the CM falls; only our intuition changes.

The derivation assumed V depends only on the separation. What breaks if there is also an external force?

The separation generally breaks — with one beautiful exception. Add an external potential $U_1(\mathbf{r}_1) + U_2(\mathbf{r}_2)$ and, rewritten in $(\mathbf{R}, \mathbf{r})$ , it couples the two coordinates: the CM no longer moves freely, and the relative motion feels the external field's gradient across the system (this is exactly how tidal forces arise). The exception is a uniform gravitational field, $U_i = -m_i\,\mathbf{g}\cdot\mathbf{r}_i$ : summed, it depends on the coordinates only through $M\mathbf{R}$ , so the CM falls with acceleration $\mathbf{g}$ while the relative motion decouples perfectly — unchanged. That is why an Earth–Moon-style analysis works even though the pair is falling around the Sun: to the (very good) extent the Sun's field is uniform across the system, the internal dance never notices the fall. The residual non-uniformity is the tide.

Why does the reduced mass shift hydrogen's spectral lines? Electrons are quantum — does this classical argument even apply?

It transfers intact, which is part of why it is so celebrated. The quantum two-body problem separates in exactly the same way: the Schrödinger equation for electron plus proton splits into a free CM part and a relative part in which the electron mass is replaced by $\mu = m_em_p/(m_e+m_p) \approx m_e(1 - 1/1836)$ . Every Bohr energy level scales with $\mu$ , so all hydrogen lines sit at frequencies a part in 1836 below the fixed-nucleus prediction. The payoff came in 1932: deuterium's nucleus is twice as heavy, its $\mu$ is slightly larger, and its spectral lines are shifted relative to ordinary hydrogen by about 1 part in 3700 — Urey resolved exactly those faint displaced lines and discovered heavy hydrogen. A classical coordinate change, surviving quantization untouched, detecting a neutron before the neutron was discovered.

Both bodies are accelerating toward each other, so how can the CM frame be inertial?

Because the accelerations are internal and cancel in the mass-weighted sum. Newton's third law makes the forces on the two bodies equal and opposite, so $m_1\ddot{\mathbf{r}}_1 + m_2\ddot{\mathbf{r}}_2 = \mathbf{F}_{12} + \mathbf{F}_{21} = 0$ , which is exactly $M\ddot{\mathbf{R}} = 0$ . The individual bodies accelerate wildly — a comet at perihelion pulls hundreds of g — but the balance point glides at constant velocity through it all, and a frame moving with it is perfectly inertial. This is worth contrasting with the frame of either body: sitting on the Earth and watching the Moon is a non-inertial view, and treating the Earth as fixed without using $\mu$ smuggles in an uncorrected fictitious force. The reduced mass is precisely the correction that makes the fixed-center fiction exact.

Which angular momentum is conserved in the two-body problem — about which point?

Two separately useful statements. About any fixed point in an inertial frame, the total angular momentum $\mathbf{L}_{\text{tot}} = M\mathbf{R}\times\dot{\mathbf{R}} + \mu\,\mathbf{r}\times\dot{\mathbf{r}}$ is conserved, because internal central forces exert no net torque. More usefully, the two pieces are conserved individually: the CM piece because $\dot{\mathbf{R}}$ is constant, and the internal piece $\mathbf{L} = \mu\,\mathbf{r}\times\dot{\mathbf{r}}$ because the relative problem is a central-force problem in its own right. It is this internal $\mathbf{L}$ — the angular momentum of the reduced particle about the force center — that pins the orbital plane and feeds every formula in the central-force articles. When astronomy texts quote "the angular momentum of the Earth–Moon system," they mean the internal piece, evaluated with $\mu$ .

When can I safely ignore the reduced-mass correction and treat the heavy body as fixed?

Estimate the fractional error first — it is simply the mass ratio $m_{\text{light}}/m_{\text{heavy}}$ — then compare with the precision you need. Earth–Sun: $3\times10^{-6}$ , so the fixed-Sun picture is fine for anything short of precision ephemerides. Moon–Earth: $1/81 \approx 1.2\%$ — already fatal for lunar theory, tide prediction, or Apollo navigation. Hydrogen: $1/1836 \approx 0.05\%$ , which sounds tiny but is enormous by spectroscopic standards (lines are measured to parts per billion). Equal-mass binaries: order 100% — the fixed-center picture is not an approximation but a different, wrong problem. The pleasant surprise is that the correction costs nothing: the exact theory is the same one-body problem with $m$ relabeled $\mu$ , so there is rarely a good reason to use the approximation at all once you know the reduction exists.