Let $(M,g)$ be a Riemannian manifold, and let $\nabla$ denote the Levi-Civita connection. Then we say a smooth curve $\gamma:J\to M, t\mapsto\gamma(t)$ is a geodesic if $$D_t\gamma'=0.$$ We say a smooth curve $\hat{\gamma}:I\to M, s\mapsto\hat{\gamma}(s)$ is a pregeodesic if there exists a diffeomorphism $\phi:J\to I$ such that $\gamma:=\hat{\gamma}\circ\phi$ is a geodesic for some open interval $J\subseteq\mathbb{R}$.

Let's now turn to the local representation of the above, that is, suppose $(U,x^j)$ are coordinates on $M$ with Christoffel symbols $\Gamma_{ij}^k$. Then then geodesic equation reads $$\ddot{\gamma}^k+\dot{\gamma}^i\dot{\gamma}^j\Gamma_{ij}^k=0.$$ Then a fairly straightforward application of the chain rule yields the result: A curve $\hat{\gamma}:I\to U$ is a pregeodesic if and only if $$\frac{d^2\hat{\gamma}^k}{ds^2}+\frac{d\hat{\gamma}^i}{ds}\frac{d\hat{\gamma}^j}{ds}\Gamma_{ij}^k=f(s)\frac{d\hat{\gamma}^k}{ds},\qquad (*)$$ for some continuous $f:I\to\mathbb{R}$. Indeed (for the relevant direction), suppose $\gamma=\hat{\gamma}\circ\phi$ is a geodesic for some diffeomorphism $\phi:J\to I, s=\phi(t)$. Then $\hat{\gamma}$ satisfies the above system $(*)$ with $$f(\phi(t))=-\frac{\frac{d^2\phi}{dt^2}}{\left(\frac{d\phi}{dt}\right)^2}.$$

This leads to my question: I've come across in the literature that $(*)$ is equivalent to the equation $$\frac{d^2\hat{\gamma}^k}{ds^2}+\frac{d\hat{\gamma}^i}{ds}\frac{d\hat{\gamma}^j}{ds}\Gamma_{ij}^k=F(\gamma')\frac{d\hat{\gamma}^k}{ds},\qquad (**)$$ for some continuous $F:TU\to\mathbb{R}$ which is homogeneous of degree $1$ in the tangent variable.

I don't understand what this function $F$ is. Clearly, any curve $\hat{\gamma}$ satisfying $(**)$ is a pregeodesic. However, the result is saying that there exists some $F:TU\to\mathbb{R}$ with the above properties so that all pregeodesics satisfy $(**)$ with that specific $F$.

Now, the function $f(s)$ depends on the diffeomorphism $\phi$, which in turn depends on the geodesic $\gamma$ with starting point $(x,\xi)\in TU$. There is some homogeneity of geodesics when dealing with the initial condition, so this certainly seems reasonable, but I can't piece all of this together in coherent form.

I think this may actually be related to general sprays, and this $F$ is a reformulation of the Liouville vector field associated to the geodesic spray, but this is a bit outside my field (at the moment).

Any help or references would be appreciated.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.