2

In the book I'm reading ("Differential Geometry Curves-Surfaces-Manifolds by Wolfgang Kuhnel") two definitions of prinicpal curvatures directions are presented:

  1. The extramum values of $II(X,X)$ under the constraint $I(X,X)=1$
  2. The eigenvectors of the shape operator $L$

I'm having some difficulty proving this equivalence. It was also mentioned that the gradient of $II(X,X)$ is simply $LX$ ($L$ operating on $X$). So I thought if I prove that, it follows that (1) holds iff the gradient of $II(X,X)$ is proportional to the gradient of $I(X,X)$ (which is clearly $X$) iff $LX$ and $X$ are proportional iff $X$ is an eigenvector.

  • How do I compute the gradient of $II(X,X)$ and show it is equal to $LX$. It seems like it is a simple task that I'm missing some idea or technique.
  • I'll be glad to hear of any interesting insights regarding prinicpal curvatures in order to have a better intuition.
stag
  • 641
  • 4
  • 14

1 Answers1

0

I suppose you could basically identify $T_uf$ as $\mathbb{R}^2$, then you could carry out the calculation in the usual way.
To be more specific about the gradient w.r.t second fundamental form:
\begin{equation} \frac {\partial}{\partial x_1}<LX,X>=<L'(X)\frac {\partial X}{\partial x_1},X> + <LX,\frac {\partial X}{\partial x_1}>=2<LX,\frac {\partial X}{\partial x_1}>=2<LX,(1,0)^t> \end{equation} which is the first coordinate of $LX$. Similarly, $$ \frac {\partial}{\partial x_2}<LX,X>=2<LX,(0,1)^t> $$ which is the second coordinate of $LX$.
Combine this two equations, you can get the result.