It is claimed in page 154 Lemma 2.49 that if there is a group homomoprhism $\varphi:\Bbb Z \rightarrow G$, then we have the following commutativity $$\require{AMScd}\begin{CD}\Bbb Z @> {\varphi}>> G \\@V{\simeq}VV@VV{\simeq}V\\ \tilde{H}_k(S^k;\Bbb Z) @>{\varphi_*}>>\tilde{H}_k(S^k;G)\end{CD}$$ by induction. How is this done?


So Hatcher makes two claims few lines earlier that are important:

  1. $\varphi_*$ commutes with homology long exact sequence
  2. $\varphi_*$ commutes with induced homomorphisms $f_*$ for any $f:(X,A)\to(Y,B)$

Both are easy to show (hopefuly). What is harder is that I think that 1. implies that $\varphi_*$ also commutes with the Mayer-Vietoris sequence. Which will be important.

Now consider the sphere $S^k$. Take $A,B$ to be the upper and lower hemisphere and apply the Mayer-Vietoris sequence to it. Since $A,B$ are contractible then there are lots of $0$s in the sequence and so we get an isomorphism $\tilde{H}_i(S^k)\simeq \tilde{H}_{i-1}(A\cap B)$ from it. And $\varphi_*$ commutes with it.

On the other hand $A\cap B$ is homotopy equivalent to $S^{k-1}$ and so $\varphi_*$ commutes with some $\tilde{H}_{i-1}(A\cap B)\simeq\tilde{H}_{i-1}(S^{k-1})$ isomorphism by property 2.

Both these facts are coefficient independent and together they show that there is a commutative diagram of the form:

$$\require{AMScd}\begin{CD}\tilde{H}_{k-1}(S^{k-1};\Bbb Z) @> {\varphi_*}>> \tilde{H}_{k-1}(S^{k-1}; G) \\@V{\simeq}VV@VV{\simeq}V\\ \tilde{H}_{k}(S^{k};\Bbb Z) @> {\varphi_*}>> \tilde{H}_{k}(S^{k}; G)\end{CD}$$

Finally by induction we reduce the problem to $k=0$. This case has to be calculated manually I'm afraid. And I'm not sure why it holds to be honest.


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.