1
$\begingroup$

Let $(M,g)$ be a Riemannian manifold of dimension $n$ and $X$ be a immersed submanifold in $M$ of dimension $k$ i.e there is a immersion $F_{0}:X \longrightarrow M$.

A deformation of the submanifold $X$ is defined by a smooth family of immersions $F : I \times X \longrightarrow M$ i.e. $F_{t}: X \longrightarrow M$ is an immersion for all $t \in I$ and $F_{0}$ is the immersion of $X$ defined above, so we have a family of immersed submanifolds $F(t,(X))=X_{t}$.

Let $F_{*}( \frac{\partial}{\partial t})$ be the deformation vector field on $M$ associated to this deformation i.e. for any $p\in X$ take the curve $F(\cdot,p):I \longrightarrow M$ and $F_{*}( \frac{\partial}{\partial t})$ at $p$ is the tangent vector to this curve at $t=0$.

Now this is my problem: I have read that if the submanifold $X$ is compact and orientable then we can find a family of diffeomorphisms of $X$ depending on $t$ such that we can assume that the vector field of the deformation is normal to $X_{t}$ for all $t$. It is not clear how this reparametrization is done, if I consider the submanifold $X_{0}$ and the deformation vector field on it, i has 2 components corresponding to the splitting $TM \vert_{X}=T(X) \oplus N(X)$ how I get rid of the tangent component using diffeomorphisms of $X$ as it is stated above?

Thanks for your help

$\endgroup$
1
  • 5
    $\begingroup$ Project $\tfrac\partial{\partial t}$ to the normal bundle and integrate. $\endgroup$ Feb 10, 2014 at 19:56

2 Answers 2

1
$\begingroup$

If you're looking for an explicit example, see page 8 in Regularity Theory for Mean Curvature Flow by Ecker. The preview on amazon includes the page.

http://www.amazon.com/Regularity-Theory-Mean-Curvature-Flow/dp/0817637818

The equation (2.1) referred to is the mean curvature flow equation, $$\frac{\partial x}{\partial t} = \vec H(x)$$

$\endgroup$
1
  • $\begingroup$ Thanks, the book you mentioned seems to be a good reference in general, very clear explanation how to reparametrize by projecting to the normal vector field (as commented by Anton and Valeri above) $\endgroup$
    – Coffee
    Feb 11, 2014 at 2:04
3
$\begingroup$

This is not exactly what you are asking, but in algebraic geometry we would say the following: the family of immersions $(F_t)$ is not uniquely defined, we can replace $F_t$ by $F_t\circ\varphi _t$, where $(\varphi _t)$ is a one-parameter family of diffeomeorphisms of $X$ -- this will give the same family of submanifolds. Thus $F_*(\frac{\partial}{\partial t} )$ is well defined only in the quotient $TM_{|X}/TX$, which is canonically isomorphic to the normal bundle.

$\endgroup$
1
  • 2
    $\begingroup$ add to abx answer: i.e., just take $\phi_t$ be the one-parameter group of diffeomorphisms of M generated by the tangent component of the deformation field. $\endgroup$
    – valeri
    Feb 10, 2014 at 18:17

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.