# Projective Space orientation

I'm trying to prove that the projective plane $\mathbb{P}^n$ is orientable is and only if $n$ is odd. To do that that, I have a hint,to prove that the antipodal map is orientation preserving if only if $n$ is odd, I've done that, but it don't know how to conclude the result.

-
How to you define $\mathbb P^n$? –  Phira May 20 '12 at 17:35
Identifying antipodal points on the sphere... –  Jr. May 20 '12 at 17:38
What do you know about the orientability of the sphere? –  Phira May 20 '12 at 17:45
I know that it is a orientable manifold... –  Jr. May 20 '12 at 17:50
So, what happens if you simply take the orientation of the sphere as orientation of the projective space? –  Phira May 20 '12 at 17:53

Let $\{O_p\}_{p \in S^n}$ be an orientation for $S^n$. If $\pi:S^n \rightarrow \mathbb P^{n}$ is the projection (a local diffeomorphism), then the idea is to define a basis $(b_1,\dots,b_n)$ of $T_q\mathbb P^{n}$ to be in $O_q'$ if $\pi_{*,p}^{-1}(b_1,\dots,b_n) \in O_p$ for any one of the two points $p$ in the fibre $\pi^{-1}(q)$.
You have already shown that this is well defined if $n$ is odd!
you mean providing a nowhere vanishing $n-$Form? –  Blah May 21 '12 at 21:16
You must show a nonwhere vanishing $n$-form: if you have $n$ tangent vectors at $q$, then you could push them up via $\pi^{-1}_{*,p}$ for the two points $p$. On $S^n$ there is a volume form - evaluate. This is well defined if $n$ is odd. (think about the pullback of the volumeform under the antipodal map) –  Blah May 22 '12 at 18:11