Why $\langle x_i,x_{jk}\rangle=0$ holds true?

89 Views Asked by At

I am reading the paper "Minimal Immersions of 2-Manifolds into Spheres". In (1.1), page 379, the author wrote that $\langle x_i,x_{jk}\rangle=0$. I don't know why this equality holds true but I think this is not a difficult one. Did I miss something? I would be grateful if someone could help me!

The link is behind a paywall, so I explain more about the notations here:

$(M,g)$ is a closed connected 2-dimensional Riemannian manifold.

Consider the isometric minimal immersions $\tilde{x}:M\rightarrow\mathbb{S}^N(1)$.

$j:\mathbb{S}^N(1)\rightarrow\mathbb{R}^{N+1}$ is the canonical embedding.

$x:=j\circ\tilde{x}$.

$x_i,x_{jk}$ is the covariant derivatives.

$\langle,\rangle$ is the inner product in $\mathbb{R}^{N+1}$.

1

There are 1 best solutions below

3
On BEST ANSWER

So nothing specific was assumed about the local coordinates. This is just the usual cyclic permutation trick for solving for Christoffel symbols. Before we start, note that the covariant derivatives $x_{ij}$ and $x_{ji}$ are equal: \begin{align*} x_{ij} &= \frac{\partial^2 x}{\partial u^j\partial u^i} + \sum\Gamma^k_{ij}\frac{\partial x}{\partial u^k} \\ &= \frac{\partial^2 x}{\partial u^i\partial u^j} + \sum\Gamma^k_{ji}\frac{\partial x}{\partial u^k} \\ &= x_{ji}, \end{align*} since the connection is torsion-free.

Since the connection is compatible with the Riemannian metric, we have $$0 = \nabla_{\partial/\partial u^k} g_{ij} = \langle x_i,x_{jk}\rangle + \langle x_{ik},x_j\rangle.$$ We cyclically permute, and do the usual $+$ $-$ $+$ trick: \begin{alignat*}2 0 &= \langle x_i,x_{jk}\rangle + \langle x_{ik},x_j\rangle & \quad (+) \\ 0 &= \langle x_j,x_{ki}\rangle + \langle x_{ji},x_k\rangle & (-) \\ 0 &= \langle x_k,x_{ki}\rangle + \langle x_{kj},x_i\rangle & (+) \end{alignat*} Using the symmetry of the covariant derivatives we asserted at the beginning, we end up with $2\langle x_i,x_{jk}\rangle = 0$, as desired.