Let $S\subset\mathbb{R}^n$ be a compact, connected, embedded hypersurface. For $x_0\notin S$, define: \begin{align*} f_{x_0}:S &\to \mathbb{S}^{n-1}\subset \mathbb{R}^n\\ x&\mapsto \frac{x-x_0}{||x-x_0||} \end{align*}

By Jordan's separation theorem, $\mathbb{R}^n\setminus S$ has two connected components $U$ (bounded) and $V$ (unbounded). Prove that the degree of $f_{x_0}$ is zero for all $x_0\in V$.

For the case $n=2$, I can easily see this is true by looking at the winding number. But I'm totally stuck with the general case.

Any ideas?


First, note that $\mathrm{deg}(f_{x_0})$ does not depend on the point $x_0 \in V$ (to see this, build an homotopy between maps with differents $x_0$'s).

Now suppose w.l.o.g that $0 \in U$. Let $x_0$ be outside of a compact big ball containing $S$. We claim that $\frac{x_0}{\|x_0\|}$ is not on the image of $f_{x_0}$. Indeed, suppose it were. Then there would be an $x \in S$ such that $$f_{x_0}(x)=\frac{x_0}{\Vert x_0\|},$$ which tells us that $$x-x_0=\Vert x-x_0\| \frac{1}{\Vert x_0\|}x_0.$$ Letting $k:=\Vert x-x_0\| \frac{1}{\Vert x_0\|}$, we see that $k>0$ and that $$x=(1+k)x_0,$$ which is an absurd, since $x$ was supposed to be of lesser length than $x_0$.

It follows that $f_{x_0}$ is not surjective. Hence, it must have degree zero.

  • $\begingroup$ I know the definitions of degree by $\int_S f^*[\omega]=\deg(f)\int_{\mathbb{S}^{n-1}}$ and by $\deg(f)=\sum_{p\in f^{-1}(q), q \text{ regular}}sign_p(f)$. How does non-surjectivity imply degree zero? $\endgroup$ – rmdmc89 Nov 25 '17 at 21:36
  • 1
    $\begingroup$ @AguirreK Using the second definition, pick a point not in the image (guaranteed to exist by surjectivity). It is automatically a regular value, and the pre-image is empty, hence the degree is zero. $\endgroup$ – Aloizio Macedo Nov 25 '17 at 21:52
  • $\begingroup$ @AguirreK Yes. With the intuition of degree being how many times the map "wraps around" the target, you can interpret that as saying that if a map misses a point then it couldn't have wrapped around the target fully, which is somewhat natural. $\endgroup$ – Aloizio Macedo Nov 25 '17 at 22:12
  • $\begingroup$ @AguirreK And sorry, I meant to write "(guaranteed to exist by lack of surjectivity)". $\endgroup$ – Aloizio Macedo Nov 25 '17 at 22:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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