MathOverflow is a question and answer site for professional mathematicians. It's 100% free, no registration required.

Sign up
Here's how it works:
  1. Anybody can ask a question
  2. Anybody can answer
  3. The best answers are voted up and rise to the top

I've been trying to understand principal bundles, and to that end have been looking at the bundle $$ \pi: SU(2) \to \mathbb{CP}^1,~~~ (a_{ij}) \mapsto [a_{11},a_{21}], $$ with fibre $U(1)$. I assumed that the bundle would be trivial over the standard nbds $U_1,U_2 \subset \mathbb{C}$, but can't seem to identify the local trivializations. Now $$ \pi^{-1}(U_1) = \{\left( \array{a & - \overline{b}\\\ b & \overline{a}} \right)|~ a \neq 0\}, ~~~ \pi^{-1}(U_2) = \{\left( \array{a & - \overline{b}\\\ b & \overline{a}} \right)|~ b \neq 0\}, $$ and any trivialization $\alpha_1:\pi^{-1}(U_i) \to U_i \times U(1)$, will map $$ \alpha_1:\left( \array{a & - \overline{b}\\\ b & \overline{a}} \right) \mapsto ([a,b],h_{a,b}^1), $$ for some $h_{a,b}^1$. Defining $h^1_{a,b} = arg(a) = \frac{a}{|a|}$, and similarly $h^2$, works, but then the transition functions are not in $U(1)$.

share|cite|improve this question
For your "similarly" is $h^2 = arg(b)$? – Ryan Budney Feb 3 '10 at 18:23
"I assumed that the bundle would be trivial over the standard nbds....". More generally, every fiber bundle trivializes over any contractible open set. Of course, this won't neccesarily help you CHOOSE you transition functions or show that everything works (which seems to be your issue), but it can certainly be helpful for later problems where you suspect something is a fiber bundle. – Jason DeVito Feb 3 '10 at 18:28
Yes, I mean $h^2_{a,b} = arg(b)$. – Dyke Acland Feb 3 '10 at 18:54
Note that this is the Hopf fibration. – Sebastian Mar 18 '10 at 7:15
up vote 3 down vote accepted

You have two charts $U_0=\Bbb C$ and $U_\infty=(\Bbb C\setminus 0)\cup \infty$. The transition function on the intersection $\Bbb C\setminus 0$ is $g(z)=z/|z|$.

share|cite|improve this answer
What about the general bundle $SU(n) \to \mathbb{CP}^{n-1}$ with fibre $U(n-1)$? What are the trivialisations $\alpha_k:\pi^{-1}(U_k) \to U_k \times U(n-1)$. My guess is that one embeds $U(n)$ into $SU(n) \times U(1)$, and then maps $(a_{ij})$ to $(h^k_{a_{ij}},\frac{a_{k1}}{|a_{k1}|})$. I 'feel' that $h^k_{a_{ij}}$ should somehow be related to the minor of $a_{i1}$ but can't see how to define it. – Dyke Acland Feb 3 '10 at 20:13
Well, there are many different ways to define these trivializations. Basically, for each of the standard charts $V_1,...,V_n$ on the projective space, choose sections $s_i: V_i\to U(n)$, and then compute the transition functions as ratios of these sections on intersections of the charts. – Pavel Etingof Feb 3 '10 at 20:30
But surely I need the trivializations to define the sections. – Dyke Acland Feb 3 '10 at 20:42
yes. So let $V_1$ be the first chart, where the 1-st coordinate is not zero. Then to a point $z=(z_1,...,z_n)$ of this chart one should attach a unitary matrix $u(z)$ with the first row proportional to $z$, and do so continuously in $z$ when $z$ is in $V_1$. This is a linear algebra problem: basically you need to construct a frame continuously depending on $z$ whose first vector is a multiple of $z$. This is not hard to do, but I don't really know the best way of doing this. And then you do similarly for $V_i$, $i>1$. – Pavel Etingof Feb 3 '10 at 20:59
Yes, this is this problem I was talking about in my earlier comment. As I said before I think the best approach is to map $V_i$ to $U(n)$ viewed as contained in $SU(n) \times U(1)$, but I'm not too sure how to do so exactly. As you said this is really a linear algebra problem (I can't see continuity being a problem) and I might re-post it as such – Dyke Acland Feb 3 '10 at 21:21

Your Answer


By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.