Assume for $\xi\in S^{n-1}$ the parametrization of a closed hypersurface is given by $x(\xi)=R(\xi)\xi\in \mathbb R^n$. Here $R: S^{n-1}\to \mathbb R$ is a positive function. Is there a reference for a proof of the formula \begin{eqnarray*} dS_{R}=R^{n-2}\sqrt{R^2+\vert\nabla R(\xi)\vert^2}\:dS_{\xi}\:? \end{eqnarray*} Notation: $dS_R$ denotes the area element on $\{R(\xi)\xi:\xi\in S^{n-1}\}$ and $dS_{\xi}$ denotes the area element on $S^{n-1}$.


1 Answer 1


Other than the chain rule, I think the only ingredient needed for this is the following formula for the determinant of a rank 1 perturbation of an invertible matrix; for $d\in \mathbb{N}$, $A\in \mathrm{GL}(d)$ and $v\in \mathbb{R}^d$,

$$ \mathrm{det}(A+vv^T) = \mathrm{det}(A)(1 + v^TA^{-1}v). $$

Writing $\hat x$ for the embedding $S^{n-1}\to \mathbb{R}^n$, the chain rule gives you the local expression

$$ \dfrac{\partial x}{\partial u^i} = R\dfrac{\partial \hat x}{\partial u^i} + \dfrac{\partial r}{\partial u^i}\hat x $$

in any chart $(U,(u^i)_{i=1}^{n-1})$ on $S^{n-1}$. Letting $g$ be the metric tensor for your hypersurface, $g$ is related to the metric tensor $\hat g$ on the sphere via

$$ g_{ij} = R^2 \hat g_{ij} + \dfrac{\partial r}{\partial u^i}\dfrac{\partial r}{\partial u^j}. $$

NB: this uses the fact that $\hat x \cdot \hat x = 1$ and $\hat x \cdot (\partial \hat x/\partial u^i)=0$.

If you write this as $g = R^2 \hat g + vv^T$ with $v$ the vector with components $\partial r/\partial u^i$, the formula above gives you

$$ \begin{array}{lll} \mathrm{det}(g) &=& \mathrm{det}(R^2 \hat g)(1 + v^T(R^2\hat g)^{-1}v) \\ &=& R^{2n-2}\left(R^2+ g^{ij}\dfrac{\partial r}{\partial u^i}\dfrac{\partial r}{\partial u^j} \right)\mathrm{det}(\hat g) \\ &=& R^{2n-2}(R^2 + |\nabla R|^2)\mathrm{det}(\hat g), \end{array} $$

which is what you want when combined with the usual expression for the volume form on a Riemannian manifold.

  • $\begingroup$ Many thanks! Indeed the determinant formula is the key to the problem. I missed that. $\endgroup$
    – guest61
    May 26, 2019 at 18:14

Your Answer

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

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