1
$\begingroup$

Let $f: M \to N$ be a diffeomorphism between two riemannian Manifolds. Suppose there exist constants $0 < c \leq C$ such that for all $p \in M$, we have $c \leq |df_p| \leq C$. Here, $df$ denotes the differential of $f$ as usual and $|df_p|$ denotes the energy density of $f$ at $p$, i.e, the norm coming from the induced riemannian metric on the bundle $TM \otimes f^*(TN)$. I wonder now if it's true that there are constants $0 < b \leq B$ with $b \leq |df^{-1}_q| \leq B$ for every $q \in N$.

$\endgroup$

1 Answer 1

1
$\begingroup$

Lower bound yes. Upper bound no.

If you take orthonormal frames for $M$ and $N$, then the norm $|df_p|$ is the Frobenius norm on the matrix expression in those frames. Note that in any bases $df_p$ and $df^{-1}_{f(p)}$ are inverse matrices. The upper bound on the Frobenius norm of $df_p$ implies that each of the singular values are bounded above by $C$, and hence their inverses are bounded below by $1/C$, which gives the lower bound.

To see that the upper bound need not hold, let $M = N = \mathbb{R}^2$ and take

$$ (u,w) = f(x,y) = (x, \sinh^{-1}(y) ) $$

For $|df|$ we have the bound $ 1 \leq |df| \leq \sqrt{2}$.

For $|df^{-1}|$ you can compute

$$ df^{-1}(u,w) = \begin{pmatrix} 1 & 0 \\ 0 & \cosh w \end{pmatrix} $$

is not bounded.


You can of course assume that $M, N$ are compact, but then the desired results are entirely trivial.

$\endgroup$
2
  • $\begingroup$ Thanks for your answer. It might still be true, though, that the upper bound holds in my specific setting. Let me elaborate: I have $M$ a non-compact riemannian manifold and some value $0 < \beta < i(M)$ and constants $0 < c \leq C$ as above, such that for each $p \in M$, the differential of the diffeomorphism $exp_p|_{B_\beta(0_p)}$ is bounded as described. $\endgroup$
    – H1ghfiv3
    Sep 11, 2015 at 6:51
  • $\begingroup$ Clearly, there then are upper and lower bounds for the differential of the inverse map on $U_\beta(p)$ (the geodesic ball about $p$ of radius $\beta$), however, can they be chosen independent of $p$ ? $\endgroup$
    – H1ghfiv3
    Sep 11, 2015 at 7:06

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.