I want to show that the non-orientable surface of genus $n$ has a 2-sheeted cover by an orientable surface of genus $n-1$. The base cases are easy: $S^2$ covers $\mathbb{RP}^2$ and I worked on a proof that the torus $\mathbb{T}^2$ covers $\mathbb{RP}^2\#\mathbb{RP}^2\approx K$ (Klein bottle).

Here are some of my constraints:

  • Nothing about the relationship between the Euler characteristic and covering maps has been proved, so I can't use this. Nor do I know anything about homology yet. Additionally, I don't yet know how to produce quotient manifolds using the techniques of group actions. The placement of this question in the text I'm using suggests that I don't need those tools. I pretty much only have knowledge of the universal covering space of manifolds.
  • I proved previously that if $M$ and $N$ are connected $n$-manifolds and $M$ has a $k$-sheeted covering space $H$, then $H\#N\#\ldots \#N$ (with $k$ copies of $N$) is a covering space for $M\#N$.

Induction using this fact seems like a good approach. Suppose $R$ is a connected sum of $n$ projective planes, an $T$ is connected sum of $n-1$ tori such that $T$ is a double cover for $R$. I want to show that $T\#\mathbb{T}^2$ is a double cover for $R\#\mathbb{RP}^2$. We know that $T\#\mathbb{RP}^2\#\mathbb{RP}^2 \approx T\#K$ is a double cover for $R\#\mathbb{RP}^2$, but I'm not sure where the go from here. We can't really cover $T\#K$ by $T\#\mathbb{T}^2$ since there are different numbers of sheets whether you're on $K$ or you're on $T$. Also, even if we could then we would end up with a quadruple cover of $R\#\mathbb{RP}^2$.

Also if anyone wants to explain to me how $\mathbb{RP}^2\#\mathbb{RP}^2\#\mathbb{RP}^2\#\mathbb{RP}^2$ is a double cover of $\mathbb{RP}^2\#\mathbb{RP}^2\#\mathbb{RP}^2$, I'd greatly appreciate it.

This type have been asked before like here and here, but my knowledge of the techniques used in the first link are scant. It's suggested that I don't need the techniques of orbit spaces or quotient manifolds to prove this claim, and I would like to develop a proof without those techniques.

  • $\begingroup$ I believe the first answer to your first link is exactly what you need... You don't have to use the Euler characteristic. You are given a description of how the $n-1$ torus is the double cover of a surface. The only thing left to show is that the latter surface is the sum of $n$ copies of the projective plane... You can do that with induction if you want... $\endgroup$ – Theo Apr 9 '15 at 23:47
  • $\begingroup$ possible duplicate of Covering space of a non-orientable surface $\endgroup$ – Theo Apr 9 '15 at 23:50
  • $\begingroup$ @Theo Please see my prompt again, I clarified what my issue with this problem. $\endgroup$ – Mnifldz Apr 10 '15 at 0:55

Let $kM$ denote the connected sum of $k$ copies of $M$.

As $3\mathbb{RP}^2 = T^2\#\mathbb{RP}^2$, we see that $(2k - 1)\mathbb{RP}^2 = (k-1)T^2\#\mathbb{RP}^2$ and $2k\mathbb{RP}^2 = (k-1)T^2\# 2\mathbb{RP}^2$.

If $n$ is odd, say $n = 2k - 1$, then $n\mathbb{RP}^2 = (2k-1)\mathbb{RP}^2 = (k-1)T^2\#\mathbb{RP}^2$. As $S^2$ double covers $\mathbb{RP}^2$, we see that $2(k - 1)T^2\# S^2 = (2k - 2)T^2 = (n - 1)T^2$ double covers $n\mathbb{RP}^2$ by the second dot point.

If $n$ is even, say $n = 2k$, then $n\mathbb{RP}^2 = 2k\mathbb{RP}^2 = (k-1)T^2\# 2\mathbb{RP}^2$. As $T^2$ double covers $2\mathbb{RP}^2$, we see that $2(k - 1)T^2\# T^2 = (2k - 1)T^2 = (n - 1)T^2$ double covers $n\mathbb{RP}^2$ by the second dot point.

As for why $4\mathbb{RP}^2$ double covers $3\mathbb{RP}^2$, note that $3\mathbb{RP}^2 = 2\mathbb{RP}^2\#\mathbb{RP}^2$ and $S^2$ double covers $\mathbb{RP}^2$, so $4\mathbb{RP}^2\# S^2 = 4\mathbb{RP}^2$ double covers $2\mathbb{RP}^2\#\mathbb{RP}^2 = 3\mathbb{RP}^2$ by the second dot point.

  • 1
    $\begingroup$ Far more concise than I could've anticipated. Thanks for the response. A clean proof is worth the wait. $\endgroup$ – Mnifldz Aug 10 '17 at 21:48
  • $\begingroup$ @PeldePinda: This is the second dot point of the question with $M = \mathbb{RP}^2$, $H = S^2$, $k = 2$, and $N = (k-1)T^2$. $\endgroup$ – Michael Albanese Apr 1 at 15:31

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.