I am searching for a proof of this fact, which is used for instance, in showing that the reflection of $S^n$ has deegre $-1$.
Rotman proves it through Mayer-Vietoris, but in my professors' notes this result is often exploited before introducing that tool.
It seems to me my notes suggest this homeomorphism comes from the "equatorial" inclusion $i:S^{n-1}\hookrightarrow S^n$ bbut I can't see why.
The same result should also hold for any suspension.

  • $\begingroup$ Do you know about the suspension isomorphism? $\endgroup$ Feb 28 '20 at 10:47
  • $\begingroup$ @BenSteffan no, but it is the kind of thing I am searching! $\endgroup$ Feb 28 '20 at 10:51
  • $\begingroup$ Suspension in this case is Mayer-Vietoris. $\endgroup$
    – Tyrone
    Feb 28 '20 at 11:22
  • $\begingroup$ @Tyrone is there a more general link between excision and Mayer Vietoris? I would be interested! $\endgroup$ Feb 28 '20 at 12:03

Consider $S^n$ to be the suspension $\Sigma S^{n - 1}$ for all $n \geq 1$. Let $C^n_+$ and $C^n_-$ denote the upper and lower cone of $\Sigma S^{n - 1}$, respectively, and let $S^{n - 1} \subseteq \Sigma S^{n - 1}$ be the equator. Let $U$ be a neighborhood of the north pole that can be excised from $C^n_+$.

We now look at the long exact sequence

$$ \begin{equation*} \cdots \rightarrow \tilde{H}_k(C^n_+) \rightarrow \tilde{H}_k(S^n) \rightarrow \tilde{H}_k(S^n, C^n_+) \rightarrow \tilde{H}_{k - 1}(C^n_+) \rightarrow \cdots \end{equation*} $$

Since $C^n_+$ is contractible, we have that $\tilde{H}_k(S^n) \cong \tilde{H}_k(S^n, C^n_+)$ for all $k$. Using our neighborhood $U$ from above, by excision, $\tilde{H}_k(S^n, C^n_+) \cong \tilde{H}_k(S^n \setminus U, C^n_+ \setminus U)$, which by (relative) deformation retraction is isomorphic to $\tilde{H}_k(C^n_-, S^{n - 1})$. Considering again the long exact sequence of the pair:

$$ \begin{equation*} \cdots \rightarrow \tilde{H}_k(S^{n - 1}) \rightarrow \tilde{H}_k(C^n_-) \rightarrow \tilde{H}_k(C^n_-, S^{n - 1}) \rightarrow \tilde{H}_{k - 1}(S^{n - 1}) \rightarrow \cdots \end{equation*} $$

we obtain that $\tilde{H}_k(C^n_-, S^{n - 1}) \cong \tilde{H}_{k - 1}(S^{n - 1})$ for all $k$, hence $\tilde{H}_k(S^n) \cong \tilde{H}_{k - 1}(S^{n - 1})$ for all $k$, and the claim follows.


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.