I'm not confident with this kind of problem, so I post my solution here to ask for checking it. I want to compute the homology groups of the space obtained from two copies of $\mathbb{R} P^2$ by gluing them along standard copies of $\mathbb{R} P^1$.

First I give it a cell structure: we know that $\mathbb{R}P^2$ has the cell structure of one $0$-cell $x$, one $1$-cell $a$ and one $2$-cell $A$ that glues to $2a$ (i.e. go around $a$ 2 times). So I believe the cell struture of our space is one $0$-cell $x$, one $1$-cell $a$ and two $2$-cells $A,B$ that both glue to $2a$. Hence we have the chain complex $$0\to \mathbb{Z}^2 \xrightarrow{d_2} \mathbb{Z} \xrightarrow{d_1}\mathbb{Z} \to 0$$ where $d_1=0$ and $d_2(A)=d_2(B)=2a$. Hence $$H_0=\mathbb{Z}/0=\mathbb{Z}$$ $$H_1=\langle a\rangle /2\langle a \rangle = \mathbb{Z}_2$$ $$H_2=\langle A-B\rangle / 0 = \mathbb{Z}$$

Is this correct?

  • 1
    $\begingroup$ Seems fine. Just be careful with your terminology: you wrote down the chain complex, not an exact sequence. The homology groups of an exact sequence are all zero! $\endgroup$ – Ethan Dlugie May 3 at 6:08
  • $\begingroup$ edited. Thank you. I am really worried about the cell structure. One move wrong and everything goes wrong. $\endgroup$ – Marcos G Neil May 3 at 6:13

What you wrote down looks right.

You can alternatively compute the homology groups by coming up with a homotopy equivalence to a simpler space. We'll use the following lemma (see for example this question for a weaker version):

Let $A, B, C$ be CW complexes, and let $f\colon A \to B$ and $g\colon A \to C$ be cellular maps. Then if $f'$ is homotopic to $f$ there is a homotopy equivalence $B\cup_{f, g} C\simeq B\cup_{f',g}C$.

Now define $D$ as the quotient $\mathbb{RP}^2 \cup_{\iota, \iota} \mathbb{RP}^2$ where $\iota\colon \mathbb{RP}^1\to \mathbb{RP}^2$ is the standard inclusion. This has a cell structure as you've described. We can think about it as starting with a copy of $\mathbb{RP}^2$ and attaching a $2$-cell via a map $\varphi\colon\partial D^2 \cong S^1\to \mathbb{RP}^1\cong S^1$ of degree $2$, i.e. $D$ is homeomorphic to $\mathbb{RP}^2\cup_{\varphi} D^2$.

However, since $\pi_1(\mathbb{RP}^2)\cong \mathbb{Z}/2$ this attaching map is null-homotopic so in fact $D$ is homotopy-equivalent to $\mathbb{RP^2}\vee S^2$ by the above lemma (cf this related question), whose homology groups are easily computable via $\tilde{H}_k(A\vee B)\cong \tilde{H}_k(A) \oplus \tilde{H}_k(B)$.

| cite | improve this answer | |
  • $\begingroup$ are you writing reduced homology at the end? THe same is true for homology groups, right? $\endgroup$ – Marcos G Neil May 3 at 16:32
  • 2
    $\begingroup$ Yes, I'm writing reduced homology at the end because the additivity statement isn't true in degree $0$. If $r_1=rank (H_0(A))$ and $r_2=rank (H_0(B))$ then $rank (H_0(A\vee B) )= r_1 + r_2 -1$ (since you're wedging together the two components containing the base points). $\endgroup$ – William May 3 at 16:35

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.