Bianchi Identities and Conservation Laws

Every field strength built as the derivative of a potential drags a hidden identity behind it. The curl of a gradient vanishes; the divergence of a curl vanishes; and in the same spirit the curvature of a connection — a kind of second derivative of that connection — cannot be arbitrary. It must satisfy differential constraints forced by the mere fact of being a curvature. Those constraints are the Bianchi identities, the quiet engine beneath the loudest conservation laws in physics. There are two. The first (algebraic) Bianchi identity is a pointwise cyclic symmetry of the Riemann tensor that holds for any torsion-free connection; it forces the pair-exchange symmetry $R_{\mu\nu\rho\sigma}=R_{\rho\sigma\mu\nu}$ and cuts the independent components of curvature in $n$ dimensions to $\tfrac1{12}n^2(n^2-1)$ — just $20$ in four dimensions. The second (differential) Bianchi identity relates covariant derivatives of the curvature, and contracting it twice with the metric yields a single symmetric tensor that is divergence-free identically: the Einstein tensor $G^{\mu\nu}=R^{\mu\nu}-\tfrac12 R g^{\mu\nu}$ , with $\nabla_\mu G^{\mu\nu}=0$ . That one line is why general relativity is consistent. Set $G_{\mu\nu}$ proportional to the stress–energy tensor and the geometry forces $\nabla_\mu T^{\mu\nu}=0$ — local energy–momentum conservation is a theorem, not a postulate. The same skeleton runs through gauge theory, where the identity collapses to $DF=0$ , and electromagnetism, where it becomes $\mathrm{d}F=0$ . This article develops the index and form pictures side by side and shows the message underneath: conservation laws are integrability conditions baked into curvature.

Key takeaways

The Riemann tensor obeys two Bianchi identities: an algebraic cyclic one, $R^{\rho}{}_{[\sigma\mu\nu]}=0$ , and a differential one, $\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0$ — both automatic for the Levi-Civita connection.
The first forces the curvature's symmetries and leaves $\tfrac1{12}n^2(n^2-1)$ independent components ( $1$ in 2D, $6$ in 3D, $20$ in 4D).
Contracting the second twice gives $\nabla_\mu G^{\mu\nu}=0$ : the Einstein tensor is divergence-free identically, off-shell, for any metric.
Through $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$ this forces $\nabla_\mu T^{\mu\nu}=0$ — energy–momentum conservation is a geometric necessity, not an assumption.
In the language of forms the identity is just $DF=0$ (or $\mathrm{d}F=0$ in electromagnetism), a one-line consequence of the nilpotency $\mathrm{d}^2=0$ .
Key idea: Curvature is a second derivative of a potential, so it obeys an integrability condition. The differential Bianchi identity $\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0$ contracts down to $\nabla_\mu G^{\mu\nu}=0$ , making the Einstein tensor divergence-free identically — and hence forcing energy–momentum conservation through the field equations.

Where this lives today

The contracted Bianchi identity is load-bearing in the most demanding physics being done. When LIGO and Virgo detect a binary black-hole merger, the waveform templates come from numerical-relativity simulations that live or die by these identities. Split spacetime into space plus time (the ADM formulation) and Einstein's ten equations separate into four constraint equations plus six evolution equations; the contracted Bianchi identity is precisely the statement that if the constraints hold on the initial slice, the evolution preserves them. For decades simulations crashed as tiny constraint violations grew exponentially; the 2005 breakthroughs (generalized-harmonic and BSSN formulations with constraint damping) tamed this Bianchi-governed instability and made the GW150914 templates possible. The same skeleton runs through particle physics and cosmology. In non-abelian gauge theory $DF=0$ forbids "magnetic" sources, so a genuine monopole must live in a topologically nontrivial bundle — the target of searches like MoEDAL at the LHC, and the reason the instanton number $\tfrac{1}{8\pi^2}\int\mathrm{tr}(F\wedge F)$ is a topological invariant. And because $\nabla_\mu g^{\mu\nu}=0$ , the identity singles out $\Lambda$ as the unique conserved vacuum-energy term — why it sits in the $\Lambda$ CDM model Planck measures, and why any modified gravity must respect a Bianchi-type identity or violate conservation.

The physical picture

Curvature is holonomy: transport a vector around a small closed loop and it returns rotated, by an amount proportional to the loop's area and the Riemann tensor of the enclosed patch. The Bianchi identities keep all these rotations mutually consistent: tile the six faces of an infinitesimal cube with loops so each interior edge is traversed once in each direction, and the edges cancel in pairs — the boundary of a boundary is empty. Translating " $\partial\partial=0$ " into curvature gives the cyclic sum of face-contributions equal to zero: geometry's version of "the net flux of a curl through a closed surface vanishes."

Figure

Curvature is the rotation a vector picks up around a small loop, proportional to R times the area. Tiling an infinitesimal cube so every interior edge cancels gives boundary-of-a-boundary = 0 — the cyclic Bianchi identity of the enclosed curvatures

The cleanest mnemonic is electromagnetism: the field strength $F_{\mu\nu}=\partial_\mu A_\nu-\partial_\nu A_\mu$ is a curvature built from $A$ , and because it is a curl, $\partial_{[\lambda}F_{\mu\nu]}=0$ automatically — the homogeneous Maxwell pair (Faraday's law and no monopoles). Gravity is the same story with a richer, non-abelian connection. The contracted version is pure economy: Einstein needed a curvature tensor automatically divergence-free to match $\nabla_\mu T^{\mu\nu}=0$ , and geometry hands him exactly one such object up to scale. The next figure traces that cascade with the numbers on it.

Figure

The contraction cascade: the second Bianchi identity contracts once to ∇^μ R_μν = ½ ∇_ν R, then to the four identities ∇_μ G^μν = 0. Coupled to matter this forces ∇_μ T^μν = 0, and it splits Einstein's ten field equations into six evolution plus four constraints

The core equations

Begin with the object every identity constrains. The Riemann tensor is quadratic in the Christoffel symbols and linear in their derivatives — a genuine "second derivative" of the metric, and the seed of every identity below.

R^{\rho}{}_{\sigma\mu\nu}=\partial_\mu\Gamma^{\rho}{}_{\nu\sigma}-\partial_\nu\Gamma^{\rho}{}_{\mu\sigma}+\Gamma^{\rho}{}_{\mu\lambda}\Gamma^{\lambda}{}_{\nu\sigma}-\Gamma^{\rho}{}_{\nu\lambda}\Gamma^{\lambda}{}_{\mu\sigma} \qquad (1)

where: $R^{\rho}{}_{\sigma\mu\nu}$ — Riemann curvature tensor, measuring the path-dependence of parallel transport; $\Gamma^{\rho}{}_{\mu\nu}$ — Christoffel symbols (Levi-Civita connection coefficients); $\partial_\mu$ — coordinate partial derivative. It reads as a curl of the connection plus a commutator — a non-abelian curl. The first Bianchi identity is a pointwise, purely algebraic statement: the total antisymmetrization of the lower three indices vanishes. It holds for any symmetric (torsion-free) connection, and it is what makes only $20$ of the naive $256$ components of $R$ in 4D independent.

R^{\rho}{}_{[\sigma\mu\nu]}=0 \quad\Longleftrightarrow\quad R^{\rho}{}_{\sigma\mu\nu}+R^{\rho}{}_{\mu\nu\sigma}+R^{\rho}{}_{\nu\sigma\mu}=0 \qquad (2)

where: $[\,\cdots]$ — total antisymmetrization over the enclosed indices. Cycle the last three indices, add, and you get zero. The second Bianchi identity is the differential one — the true engine, antisymmetrizing a covariant derivative of the curvature over three indices and vanishing identically for the Levi-Civita connection.

\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0 \quad\Longleftrightarrow\quad \nabla_\lambda R^{\kappa}{}_{\rho\mu\nu}+\nabla_\mu R^{\kappa}{}_{\rho\nu\lambda}+\nabla_\nu R^{\kappa}{}_{\rho\lambda\mu}=0 \qquad (3)

where: $\nabla_\lambda$ — covariant derivative (metric-compatible, torsion-free); $R_{\mu\nu\rho\sigma}$ — fully-lowered Riemann tensor. The cyclic sum of covariant derivatives of curvature vanishes — the exact analogue of $\partial_{[\lambda}F_{\mu\nu]}=0$ . To extract conservation you must contract, which needs the two traces of Riemann: the Ricci tensor and the Ricci scalar.

R_{\mu\nu}=R^{\lambda}{}_{\mu\lambda\nu},\qquad R=g^{\mu\nu}R_{\mu\nu} \qquad (4)

where: $R_{\mu\nu}$ — Ricci tensor (the one independent trace of Riemann); $R$ — Ricci scalar; $g^{\mu\nu}$ — inverse metric that raises and contracts indices. In 4D the Ricci tensor is symmetric with $10$ components, one for each independent Einstein equation. Contracting the second Bianchi identity twice produces the most consequential line in the subject: a symmetric tensor whose covariant divergence vanishes for every metric.

\nabla_\mu G^{\mu\nu}=0,\qquad G^{\mu\nu}\equiv R^{\mu\nu}-\tfrac12 R\,g^{\mu\nu} \qquad (5)

where: $G^{\mu\nu}$ — Einstein tensor, the unique symmetric, divergence-free curvature combination linear in second derivatives of the metric. Geometry is self-conserving: the divergence is zero as an identity, off-shell. Couple this geometry to matter and the divergence-free property propagates onto the source — conservation delivered by pure geometry.

G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}\ \Longrightarrow\ \nabla_\mu T^{\mu\nu}=0 \qquad (6)

where: $\Lambda$ — cosmological constant (covariantly constant, since $\nabla_\mu g^{\mu\nu}=0$ ); $G$ — Newton's constant; $T_{\mu\nu}$ — stress–energy tensor. Take the divergence of both sides: the left gives zero identically, so the right must too. Finally, the whole story compresses into one line in the language of forms — the connection a Lie-algebra-valued 1-form $\omega$ , the curvature a 2-form $F$ , and the identity the vanishing of the exterior covariant derivative of $F$ .

F=\mathrm{d}\omega+\omega\wedge\omega,\qquad DF\equiv\mathrm{d}F+\omega\wedge F-F\wedge\omega=0 \qquad (7)

where: $\omega$ — connection 1-form; $F$ — curvature 2-form; $D$ — exterior covariant derivative. In electromagnetism the group is abelian and this collapses to $\mathrm{d}F=0$ — the homogeneous Maxwell equations $\partial_{[\lambda}F_{\mu\nu]}=0$ .

Derivation AThe second Bianchi identity from a normal-coordinate calculation

The goal is to prove the differential identity $\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0$ and show why it is inevitable. The strategy exploits a freedom of the Levi-Civita connection: at any point we may choose Riemann normal coordinates in which the Christoffel symbols vanish, though their derivatives do not. There the identity reduces to the elementary fact that mixed partial derivatives commute, which tensor character then promotes everywhere.

Step 1. At an arbitrary point $p$ choose Riemann normal coordinates so $\Gamma^{\rho}{}_{\mu\nu}(p)=0$ ; the quadratic $\Gamma\Gamma$ terms in equation (1) drop out, leaving $R^{\kappa}{}_{\rho\mu\nu}(p)=\partial_\mu\Gamma^{\kappa}{}_{\nu\rho}-\partial_\nu\Gamma^{\kappa}{}_{\mu\rho}$ . Why: normal coordinates, always available for a torsion-free connection, strip the curvature to the part that carries the identity.

Step 2. Differentiate. The extra connection terms inside $\nabla_\lambda$ each multiply $\Gamma(p)=0$ , so at $p$ the covariant derivative is the ordinary one: $\nabla_\lambda R^{\kappa}{}_{\rho\mu\nu}(p)=\partial_\lambda\partial_\mu\Gamma^{\kappa}{}_{\nu\rho}-\partial_\lambda\partial_\nu\Gamma^{\kappa}{}_{\mu\rho}$ . Why: $\Gamma(p)=0$ but $\partial\Gamma\neq0$ is the loophole that lets a covariant object be computed with elementary calculus.

Step 3. Form the cyclic sum over $(\lambda,\mu,\nu)$ . Every term has the shape $\partial_a\partial_b\Gamma^{\kappa}{}_{c\rho}$ , and since $\partial_a\partial_b=\partial_b\partial_a$ the six terms cancel in pairs — hence $\nabla_\lambda R^{\kappa}{}_{\rho\mu\nu}+\nabla_\mu R^{\kappa}{}_{\rho\nu\lambda}+\nabla_\nu R^{\kappa}{}_{\rho\lambda\mu}=0$ at $p$ . Why: the identity is nothing deeper than the symmetry of mixed partials in tensor dress — which is why it holds off-shell, for any metric.

Step 4. Promote to a tensor equation. The left side is a tensor; vanishing at $p$ in one frame, it vanishes at $p$ in every frame, and since $p$ was arbitrary it vanishes everywhere. Why: this tensor promotion upgrades a local coordinate trick into a universal law. *Verified: matches Wald eq. (3.2.13) and MTW §15.3 under the sign convention of equation (1).*

Go deeper · Derivation B · 1 minContracting twice to the divergence-free Einstein tensor

Derivation BContracting twice to the divergence-free Einstein tensor

The goal is to squeeze the four-index differential identity down to a two-index conservation law, seeing where antisymmetry, metric compatibility, and the Ricci traces enter. We contract the second Bianchi identity once to relate the Ricci divergence to the gradient of the scalar curvature, contract again, and rearrange to expose the Einstein tensor — every step an algebraic identity valid for an arbitrary metric, with no field equation used.

Step 1. Take the second Bianchi identity and contract $\kappa$ with $\mu$ . Using $R^{\mu}{}_{\rho\mu\nu}=R_{\rho\nu}$ and the antisymmetry $R^{\mu}{}_{\rho\lambda\mu}=-R_{\rho\lambda}$ gives $\nabla_\lambda R_{\rho\nu}-\nabla_\mu R^{\mu}{}_{\rho\nu\lambda}-\nabla_\nu R_{\rho\lambda}=0$ . Why: the first contraction turns two Riemann indices into a Ricci tensor, leaving a statement about Ricci and one surviving Riemann term.

Step 2. Contract again with $g^{\rho\nu}$ . Metric compatibility, $\nabla_\lambda g^{\rho\nu}=0$ , lets the metric slide through $\nabla$ , so $g^{\rho\nu}\nabla_\lambda R_{\rho\nu}=\nabla_\lambda R$ and the surviving Riemann term becomes another Ricci divergence. The result collapses to $\nabla_\lambda R-2\nabla_\mu R^{\mu}{}_{\lambda}=0$ , i.e. $\nabla^\mu R_{\mu\lambda}=\tfrac12\nabla_\lambda R$ . Why: metric compatibility is the linchpin — without it the metric cannot pass through $\nabla$ and the Ricci scalar never assembles.

Step 3. Rearrange into a single divergence. Write $\nabla_\lambda R=\nabla_\mu(\delta^{\mu}{}_{\lambda}R)$ and move half of it inside: $\nabla_\mu\!\left(R^{\mu}{}_{\lambda}-\tfrac12\delta^{\mu}{}_{\lambda}R\right)=0$ . Raising $\lambda$ gives $\nabla_\mu G^{\mu\nu}=0$ with $G^{\mu\nu}=R^{\mu\nu}-\tfrac12 R g^{\mu\nu}$ . Why: the one-half is forced by the "2" of the double contraction — which is why $G^{\mu\nu}$ , not $R^{\mu\nu}$ alone, is the conserved object.

Step 4. Only antisymmetry, the Ricci definitions, and metric compatibility entered — never a field equation. So $\nabla_\mu G^{\mu\nu}=0$ is off-shell, and coupling $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G\,T_{\mu\nu}$ forces $\nabla_\mu T^{\mu\nu}=0$ (since $\nabla_\mu(\Lambda g^{\mu\nu})=0$ ). Why: conservation of the source is inherited from an identity geometry satisfies regardless of dynamics. *Verified: reproduces $\nabla^\mu R_{\mu\nu}=\tfrac12\nabla_\nu R$ and $\nabla_\mu G^{\mu\nu}=0$ ; consistent with Wald ch. 3 and Lovelock's theorem (in 4D the only identically-conserved symmetric tensors algebraic in curvature are $G_{\mu\nu}$ and $g_{\mu\nu}$ ).*

Go deeper · Derivation C · 1 minThe Bianchi identity as d squared equals zero in the language of forms

Derivation CThe Bianchi identity as d squared equals zero in the language of forms

The goal is to show that the differential Bianchi identity — in every gauge theory including gravity — is a one-line consequence of the nilpotency $\mathrm{d}^2=0$ , and to read off both the homogeneous Maxwell equations and charge conservation from the same source. We differentiate the curvature 2-form's definition, watch the cubic terms cancel, then specialize to electromagnetism, where applying $\mathrm{d}$ to the source equation forces charge continuity.

Step 1. Apply the exterior derivative to $F=\mathrm{d}\omega+\omega\wedge\omega$ . With $\mathrm{d}^2\omega=0$ and the graded Leibniz rule, $\mathrm{d}F=\mathrm{d}\omega\wedge\omega-\omega\wedge\mathrm{d}\omega$ ; substituting $\mathrm{d}\omega=F-\omega\wedge\omega$ the cubic $\omega^{\wedge3}$ terms cancel, leaving $\mathrm{d}F=F\wedge\omega-\omega\wedge F=-[\omega,F]$ , i.e. $DF\equiv\mathrm{d}F+[\omega,F]=0$ . Why: the nilpotency $\mathrm{d}^2=0$ is the same "boundary of a boundary is empty" that killed the cube's interior edges, and $D$ is built exactly so the cancellation reads as "the covariant derivative of the curvature vanishes."

Step 2. Specialize to electromagnetism, where the group is abelian and $DF=0$ becomes $\mathrm{d}F=0$ , i.e. $\partial_{[\lambda}F_{\mu\nu]}=0$ — the homogeneous Maxwell pair ( $\nabla\cdot\mathbf B=0$ and Faraday's law), automatic because $F=\mathrm{d}A$ . Applying $\mathrm{d}$ to the source equation $\mathrm{d}\star F=\star J$ with $\mathrm{d}^2=0$ gives $\mathrm{d}\star J=0$ , i.e. $\partial_\mu J^\mu=0$ — charge conservation. Why: no monopoles, Faraday, and charge conservation are all kinematic shadows of the same $\mathrm{d}^2=0$ , not dynamical inputs. *Verified: $\partial_\nu\partial_\mu F^{\mu\nu}=0=\partial_\nu J^\nu$ holds identically for antisymmetric $F$ ; matches Jackson and Nakahara ch. 10.*

Limiting cases

The Bianchi identities specialize sharply across settings, each limit revealing which ingredient does the work.

Case	Result	Meaning
Flat space	$R^{\rho}{}_{\sigma\mu\nu}=0$ , identities vacuous	Both identities are trivially satisfied; $\nabla_\mu T^{\mu\nu}=0$ collapses to the ordinary $\partial_\mu T^{\mu\nu}=0$ of special relativity.
Electromagnetism (abelian)	$\mathrm{d}F=0 \iff \partial_{[\lambda}F_{\mu\nu]}=0$	The commutator drops out; the identity is the homogeneous Maxwell pair, distinct from the sourced $\mathrm{d}\star F=\star J$ .
Connection with torsion	$R^{\rho}{}_{[\sigma\mu\nu]}=\nabla_{[\sigma}T^{\rho}{}_{\mu\nu]}+T^{\lambda}{}_{[\sigma\mu}T^{\rho}{}_{\nu]\lambda}$	The first identity picks up torsion terms; the clean cyclic form needs a torsion-free connection.
Cosmological constant	$\nabla_\mu\!\left(R^{\mu\nu}-\tfrac12 R g^{\mu\nu}+\Lambda g^{\mu\nu}\right)=0$	Since $\nabla_\mu g^{\mu\nu}=0$ , a constant $\Lambda$ is the only conserved term one may add — Bianchi fixes the vacuum-energy form.
Maximally symmetric (de Sitter)	$R_{\mu\nu}=\tfrac{R}{4}g_{\mu\nu}$ , $R=\text{const}$	The contracted identity forces $R$ constant: $\nabla_\mu G^{\mu\nu}=0$ demands $\nabla_\nu R=0$ .

Worked examples

1 · Counting the independent components of curvature (confidence builder)

A rank-4 tensor in $n$ dimensions has $n^4$ components. Use the symmetries — including the first Bianchi identity — to show the Riemann tensor has only $\tfrac1{12}n^2(n^2-1)$ independent components, and evaluate for $n=2,3,4$ .

Step 1. Antisymmetry in each pair of $R_{\mu\nu\rho\sigma}$ leaves $N=\tfrac12 n(n-1)$ independent values per pair, and pair-exchange symmetry $R_{\mu\nu\rho\sigma}=R_{\rho\sigma\mu\nu}$ makes $R$ a symmetric matrix in the two composite indices — $\tfrac12 N(N+1)$ components.

Step 2. Impose the first Bianchi identity: the totally antisymmetric part $R_{[\mu\nu\rho\sigma]}$ must vanish, one condition for each set of four distinct indices, i.e. $\binom{n}{4}$ . Subtracting, $\tfrac12 N(N+1)-\binom{n}{4}$ collapses (with $N=\tfrac12 n(n-1)$ ) to $\dfrac{n^2(n^2-1)}{12}$ .

Step 3. Evaluate: $n=2$ gives $\dfrac{4\cdot 3}{12}=\mathbf{1}$ (the Gaussian curvature); $n=3$ gives $\dfrac{9\cdot 8}{12}=\mathbf{6}$ (Riemann fully fixed by Ricci); and $n=4$ gives $\dfrac{16\cdot 15}{12}=\mathbf{20}$ .

Step 4. The 4D count splits as $10$ Ricci (fixed locally by matter) plus $10$ Weyl (the free gravitational field — tidal forces and gravitational waves in vacuum). In 4D: 20 independent Riemann components = 10 Ricci + 10 Weyl, and the first Bianchi identity supplies the $\binom{4}{4}=1$ condition that trims the naive count to 20.

2 · The cosmological continuity equation from the contracted Bianchi identity

In an FRW universe with a perfect fluid $T^{\mu}{}_{\nu}=\mathrm{diag}(-\rho,p,p,p)$ and scale factor $a(t)$ , the contracted Bianchi identity forces $\nabla_\mu T^{\mu\nu}=0$ . Work out what that conservation law says, and specialize it to matter, radiation, and vacuum.

Step 1. The $\nu=0$ (time) component of $\nabla_\mu T^{\mu\nu}=0$ is $\partial_\mu T^{\mu}{}_{0}+\Gamma^{\mu}{}_{\mu\lambda}T^{\lambda}{}_{0}-\Gamma^{\lambda}{}_{\mu 0}T^{\mu}{}_{\lambda}$ . In FRW the relevant Christoffels are $\Gamma^{i}{}_{j0}=\dfrac{\dot a}{a}\delta^{i}{}_{j}$ and $\Gamma^{\mu}{}_{\mu 0}=3\dfrac{\dot a}{a}$ , with $H\equiv\dot a/a$ the Hubble rate.

Step 2. Substitute the diagonal fluid: the time derivative gives $-\dot\rho$ , the trace term $-3H\rho$ , and the last term $-3Hp$ (three spatial pressures each weighted by $H$ ). Collecting, $-\dot\rho-3H\rho-3Hp=0$ , i.e.

\dot\rho+3H(\rho+p)=0 \qquad (\star)

the cosmological continuity equation — the fluid form of energy–momentum conservation, delivered entirely by the contracted Bianchi identity.

Step 3. With an equation of state $p=w\rho$ , $\dot\rho/\rho=-3(1+w)\,\dot a/a$ integrates to $\rho\propto a^{-3(1+w)}$ . For pressureless matter ( $w=0$ ): $\rho\propto a^{-3}$ (dilution by volume). For radiation ( $w=\tfrac13$ ): $\rho\propto a^{-4}$ (an extra $a^{-1}$ from redshift). For a cosmological constant ( $w=-1$ ): $\dot\rho=0$ , a density that does not dilute at all.

Step 4. These three scaling laws of $\Lambda$ CDM cosmology — matter $\rho\propto a^{-3}$ , radiation $a^{-4}$ , vacuum const — are each a direct consequence of $\nabla_\mu G^{\mu\nu}=0$ through the contracted Bianchi identity.

Physics of the portrait

The identities carry the name of an Italian geometer who systematized nineteenth-century differential geometry — though others glimpsed the result first, its physical destiny decades ahead.

Origins · 1 minLuigi Bianchi and how this was discovered

Luigi Bianchi

1856–1928 · differential geometry and the identities of curvature

Luigi Bianchi was a leading figure of the Italian school of differential geometry that flowered around Pisa in the late nineteenth century. Born in Parma in 1856, he studied at the Scuola Normale Superiore under Betti and Dini and spent his whole career at Pisa, eventually directing the Scuola Normale. He published the differential identities now bearing his name in 1902, in work on the four-index symbols of Riemann; characteristically, they had been written down earlier by Aurel Voss in 1880 and were known to Gregorio Ricci-Curbastro, whose absolute differential calculus (the tensor calculus, with Levi-Civita) gave them their natural home. His other lasting gift to physics is the classification of the real three-dimensional Lie algebras — the nine Bianchi types — the organizing scheme for all spatially homogeneous cosmologies, so anisotropic universes are still labelled Bianchi I through IX. He never saw the deepest use of his identities: it was Einstein, Hilbert, and their successors who realized that the contracted identity is exactly what makes general relativity consistent with the conservation of energy and momentum. Bianchi died in Pisa in 1928, a geometer whose quiet identities turned out to hold the universe together.

History

1854

— Riemann introduces the curvature tensor in his Habilitation lecture on the foundations of geometry

1880

— Aurel Voss writes down the differential curvature identities later named for Bianchi

1902

— Bianchi publishes the second (differential) identity in his study of the four-index symbols

1915

— Einstein's field equations demand a curvature tensor that is identically conserved

1954

— Yang and Mills carry curvature and the identity DF = 0 into non-abelian gauge theory

Going further

Texts. Misner, Thorne and Wheeler, Gravitation, §§13–15 · Wald, General Relativity, ch. 3–4 · Nakahara, Geometry, Topology and Physics, ch. 7 and 10 (connections, curvature, and $DF=0$ on fibre bundles) · Stone and Goldbart, Mathematics for Physics, differential-geometry chapters · Arfken, Weber and Harris, Mathematical Methods for Physicists, 7th ed., tensor chapter · Hassani, Mathematical Physics, differential-geometry part · Byron and Fuller, Mathematics of Classical and Quantum Physics, tensor chapter.

Provenance. Equations are standard from Wald (ch. 3), MTW (§§13–15), and Nakahara (ch. 7). A verifier should: (1) rederive $R^{\rho}{}_{[\sigma\mu\nu]}=0$ from the symmetry of second covariant derivatives of a scalar; (2) reproduce $\nabla^\mu R_{\mu\nu}=\tfrac12\nabla_\nu R$ and $\nabla_\mu G^{\mu\nu}=0$ ; (3) verify $DF=0$ from $\mathrm{d}^2=0$ ; (4) check $\tfrac1{12}n^2(n^2-1)=1,6,20$ for $n=2,3,4$ ; (5) rederive $\dot\rho+3H(\rho+p)=0$ with scalings $a^{-3}$ , $a^{-4}$ , const. No empirical numbers are asserted; $8\pi G$ uses geometrized units $c=1$ .

Prerequisites. The Covariant Derivative · Christoffel Symbols and Affine Connections · Tensor Contraction and the Trace · Exterior Calculus · Raising and Lowering Indices. Leads to. The Metric Tensor · Hodge Star and the Codifferential · Differential Forms.

Ask the tutor: Ask me to walk through the double contraction of the second Bianchi identity line by line, or to show exactly how $\nabla_\mu G^{\mu\nu}=0$ forces the cosmological continuity equation in your favourite spacetime.

What is the difference between the first and second Bianchi identities?

They live at different levels. The first (algebraic) identity, $R^{\rho}{}_{[\sigma\mu\nu]}=0$ , is a pointwise symmetry with no derivatives; it cuts down the independent components of $R$ and forces the pair-exchange symmetry. The second (differential) identity, $\nabla_{[\lambda}R_{\mu\nu]\rho\sigma}=0$ , relates covariant derivatives of the curvature and is the one that contracts into conservation laws. The first is about the shape of curvature, the second about how it changes — both automatic for the torsion-free Levi-Civita connection.

Why is the factor of one-half in the Einstein tensor not adjustable?

Because it is fixed by the double contraction, not chosen by hand. Contracting the second Bianchi identity twice gives $\nabla^\mu R_{\mu\nu}=\tfrac12\nabla_\nu R$ , where the "2" comes from the two Ricci terms that combine. The only way to package this as a single vanishing divergence is $\nabla_\mu(R^{\mu}{}_\nu-\tfrac12\delta^{\mu}{}_\nu R)=0$ ; any other coefficient $\alpha\neq\tfrac12$ leaves a nonzero divergence $\propto(\tfrac12-\alpha)\nabla^\nu R$ . So $G_{\mu\nu}=R_{\mu\nu}-\tfrac12 R g_{\mu\nu}$ is the unique divergence-free combination — geometry left Einstein no freedom in that $\tfrac12$ .

How can conservation of energy be a theorem rather than an assumption?

In pre-relativistic physics you impose $\partial_\mu T^{\mu\nu}=0$ as a separate law. In general relativity you cannot: the field equation sets $G_{\mu\nu}+\Lambda g_{\mu\nu}=8\pi G T_{\mu\nu}$ , whose left side has identically zero covariant divergence by the contracted Bianchi identity. Take the divergence of both sides — the left gives $0$ automatically, so the right must too, and $\nabla_\mu T^{\mu\nu}=0$ is forced with no extra postulate. It is the local face of Noether's theorem for the diffeomorphism symmetry of gravity.

Does the Bianchi identity forbid magnetic monopoles entirely?

Locally yes, globally no. Wherever the field strength is exact, $F=\mathrm{d}A$ , the identity $\mathrm{d}F=\mathrm{d}^2 A=0$ holds automatically, and its component $\nabla\cdot\mathbf B=0$ says there is no magnetic charge there. A monopole evades this not by breaking the identity at a point but by living in a topologically nontrivial bundle where no single global $A$ exists — Dirac's "string," and the modern statement that its magnetic charge is a Chern number of the bundle. So monopoles are permitted only as global, topological objects, and their absence in ordinary electromagnetism means its bundle is trivial.

Where does torsion change the story?

The first Bianchi identity is the sensitive one. Its clean cyclic form $R^{\rho}{}_{[\sigma\mu\nu]}=0$ needs a torsion-free connection, because the underlying step — that second covariant derivatives of a scalar are symmetric — fails with torsion: $\nabla_{[\mu}\nabla_{\nu]}f=-\tfrac12 T^{\lambda}{}_{\mu\nu}\nabla_\lambda f\neq0$ . The identity then picks up source terms $R^{\rho}{}_{[\sigma\mu\nu]}=\nabla_{[\sigma}T^{\rho}{}_{\mu\nu]}+T^{\lambda}{}_{[\sigma\mu}T^{\rho}{}_{\nu]\lambda}$ , as in Einstein–Cartan gravity where torsion couples to spin. General relativity is torsion-free by construction, so the textbook identities apply unmodified.

Why is 3D gravity special according to the component count?

Because in three dimensions the Riemann tensor has $6$ independent components — and so does the Ricci tensor, so Riemann is algebraically determined by Ricci (the Weyl tensor vanishes in 3D). Since Einstein's equations fix Ricci from the local matter, in 3D they fix the entire curvature: nothing is left over to propagate through empty space. So 3D gravity has no local gravitational waves and no Newtonian force between static masses. The jump to $20$ components in 4D opens the $10$ free Weyl components — tidal forces and gravitational radiation.