advanced30 min studyLesson 455 in Mathematics

Jacobi fields and geodesic variation

Differentiate a family of geodesics, retain covariant derivatives and the curvature sign, and locate conjugate points in constant curvature.

On this page

What you’ll understand

  • Derive the Jacobi equation from a geodesic variation.
  • Solve the normal scalar equation in constant sectional curvature.
  • Distinguish a conjugate point from general endpoint or cut-point nonuniqueness.

A single geodesic describes one route; a Jacobi field describes the first-order separation of nearby geodesic routes. Curvature enters because differentiating a geodesic equation across a moving family does not commute freely with differentiation along the route. The resulting linear equation connects local curvature to focusing, singularities of the exponential map, and limits of distance minimization, provided its initial conditions and covariant derivatives are handled correctly.

Differentiate the family, not a coordinate sketch

Let F(s,t) be a smooth variation whose t-curves are affinely parameterized geodesics. Along the central curve write T for its velocity and J for the s-variation field. The coordinate fields in the parameter rectangle commute, and torsion freedom gives ∇sT=∇tJ\nabla_sT=\nabla_tJ. Differentiate the geodesic identity ∇tT=0\nabla_tT=0 in the s-direction and commute the covariant derivatives using the curvature convention from the preceding lesson. This produces the Jacobi equation displayed below.

The derivative D/dt acts along the moving tangent spaces; replacing it by ordinary coordinate differentiation omits connection terms. Initial values J(0) and DJ/dt(0) determine a unique solution of this linear second-order equation along a given geodesic. Conversely, locally every Jacobi field can be realized by varying initial geodesic data. Tangential fields of the form (a+bt)T arise from changing the starting parameter or speed, so not every Jacobi field represents normal spatial focusing.

D2Jdt2+R(J,T)T=0,R(X,Y)Z=∇X∇YZ−∇Y∇XZ−∇[X,Y]Z\frac{D^2J}{dt^2}+R(J,T)T=0,\qquad R(X,Y)Z=\nabla_X\nabla_YZ-\nabla_Y\nabla_XZ-\nabla_{[X,Y]}Z
The operator convention is repeated to fix the sign. T is the velocity of an affinely parameterized geodesic; the two derivatives of J are covariant.↔ Scroll wide equations horizontally.

A normal field reduces to a scalar model

On a unit-speed geodesic in a constant-sectional-curvature space of dimension at least two, choose a unit parallel normal field E. A field J=jE then satisfies the scalar equation j′′+Kj=0j''+Kj=0. With initial conditions j(0)=0 and j'(0)=1, positive curvature gives a sine, zero curvature gives linear separation, and negative curvature gives hyperbolic sine. The normalization of initial derivative is important: changing it scales the entire field and changes its interpretation as a variation of initial direction.

These formulas are exact model-space calculations, not a replacement for a general curvature-varying matrix equation. The parallel normal frame also matters; a rotating frame introduces extra derivative terms. In positive curvature the sine first returns to zero at π divided by the square root of K. This is infinitesimal refocusing of distinct initial directions. In negative curvature, the chosen normal solution grows faster than linear and has no second zero after the initial point.

j(t)={sin⁡(K t)/K,K>0,t,K=0,sinh⁡(−K t)/−K,K<0,j(0)=0, j′(0)=1j(t)=\begin{cases}\sin(\sqrt K\,t)/\sqrt K,&K>0,\\t,&K=0,\\\sinh(\sqrt{-K}\,t)/\sqrt{-K},&K<0,\end{cases}\qquad j(0)=0,\ j'(0)=1
The scalar formula assumes unit speed, constant sectional curvature, and a parallel unit normal direction. It does not describe the tangential Jacobi solutions.↔ Scroll wide equations horizontally.

Conjugacy is a derivative-level obstruction

A point at positive parameter time along a geodesic is conjugate to its starting point if a nonzero Jacobi field vanishes at both times. For variations fixing the starting point, this corresponds to a nontrivial kernel in the differential of the exponential map at the appropriate initial velocity. Conjugacy concerns infinitesimal initial-data degeneracy. Distinct finite geodesic routes can also meet without such a differential singularity, so endpoint nonuniqueness is a broader phenomenon.

The second variation relates Jacobi fields to minimizing behavior: a geodesic segment cannot remain minimizing after an interior conjugate point. At an endpoint conjugate point, it may still minimize, as an antipodal great-circle arc on a sphere does. A cut point can occur earlier because a competing minimizing route appears. The flat cylinder later supplies exactly that distinction: its topology can create a cut point even though its zero-curvature normal Jacobi fields do not refocus.

Focusing on a sphere of radius two

A round two-sphere of radius two has K=1/4. Along a unit-speed geodesic, find the normal Jacobi field with J(0)=0 and DJ/dt(0) a unit normal. Locate its first positive zero.

  1. Choose a parallel unit normal direction E along the geodesic. The scalar coefficient satisfies j''+(1/4)j=0, j(0)=0, and j'(0)=1.

  2. The solution is j(t)=2sin⁡(t/2)j(t)=2\sin(t/2). Its derivative is cos⁡(t/2)\cos(t/2), which verifies the unit initial derivative; its second derivative is −(1/2)sin⁡(t/2)-(1/2)\sin(t/2), canceling (1/4)j.

  3. The first positive zero is t=2π. This equals the distance to the antipode on a sphere of radius two. The arc to that endpoint can still minimize; continuing beyond it loses minimizing behavior.

The result

The field is 2sin(t/2)E(t), with its first conjugate point at geodesic distance 2π.

Explore the model

Interactive lab

Interactive lab · oscillator

Recover the correct initial condition from velocity

Pause and Reset, then set damping ζ = 0 and frequency ω₀ = 1. The instrument starts x(0)=1,v(0)=0, so its displacement is not the Jacobi solution with j(0)=0,j′(0)=1. Instead compute j=−v/ω₀². After two Steps, t=1, predict j=sin(1)≈0.8415 from the velocity readout. Reset and set both ζ = 0 and ω₀ = 2; at t=1 predict j=sin(2)/2≈0.4546. This reconstructs the positive constant-curvature scalar model, not general geodesic separation.

Recover the correct initial condition from velocityUnit mass, x(0) = 1 and v(0) = 0. The analytic solution includes underdamped, critically damped and overdamped cases; time is measured in seconds. Current readings: Displacement x(t): 1; Velocity v(t): 0; Regime: underdamped. All parameters are adjustable with the labeled controls below.02.557.510−1.15−0.5700.571.15Time · sDisplacementunderdamped
Displacement x(t)Current time
Displacement x(t)
1
Velocity v(t)
0
Regime
underdamped
t = 0 s

Unit mass, x(0) = 1 and v(0) = 0. The analytic solution includes underdamped, critically damped and overdamped cases; time is measured in seconds. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forMatching an equation is insufficient unless the initial data and differentiated quantity match too; the cosine displacement must be converted to the normalized sine solution.

Check your understanding

With the module's curvature convention, which equation governs a Jacobi field along an affinely parameterized geodesic?

Try it yourself

Solve the same normalized normal Jacobi initial-value problem in constant curvature K=−4. Verify j'(0), determine whether it has another positive zero, and compare it with the flat solution for t>0.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

Here √(−K)=2. Differentiate the hyperbolic sine formula, and use sinh(u)>u for positive u.

Compare with the solution

The solution is j(t)=sinh⁡(2t)/2j(t)=\sinh(2t)/2, with derivative j′(t)=cosh⁡(2t)j'(t)=\cosh(2t) and second derivative 2sinh⁡(2t)=4j(t)2\sinh(2t)=4j(t). Thus it satisfies j''−4j=0 and j'(0)=1. Hyperbolic sine is positive for t>0, so there is no second positive zero. Also sinh(2t)>2t, giving j(t)>t, larger than the flat solution with the same initial derivative. This concerns the selected unit-speed normal model, not every arbitrary Jacobi initial condition.

Take it with you

  • Jacobi fields linearize geodesic initial data using covariant derivatives.
  • The sign and initial normalization determine sine, linear, or hyperbolic-sine separation.
  • Conjugate points are exponential-map derivative singularities; cut points can have other causes.

Read further

Original sources and deeper explanations to follow at your own pace.

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.