The inclusion $U(1)\times U(n-1)\to U(n)$ induces a map $BU(1)\times BU(n-1)\to BU(n)$. This source claims that the fiber of this map is $\mathbb{C}P^n$. Could someone please explain why this is true?


There is a typo --- it should be $\mathbb{C}P^{n-1}$ (as you can also see from the cohomology ring in the next sentence).

$BU(1)\times BU(n-1)\to BU(n)$ is the classifying map for taking orthogonal direct sum of a rank-$1$ bundle and a rank-$(n-1)$ bundle. So the fibers are ways to split an $n$-dimensional vector space into the sum of $1$- and $(n-1)$-dimensional subspaces, but that is just choosing $1$-dimensional subspaces from a fixed $n$-dimensionsional one (the $n-1$ must then be its orthogonal complement), which is $\mathbb{C}P^{n-1}$.

| cite | improve this answer | |
  • $\begingroup$ Ah, got it. Thanks. $\endgroup$ – Judson Kuhrman Jul 26 at 2:50
  • $\begingroup$ A follow up question: how do we know Leray-Hirsch applies here? i.e. Why is the generator of the cohomology of $\mathbb{C}P^{n−1}$ the pullback of some element of the cohomology of $BU(1)×BU(n−1)$? $\endgroup$ – Judson Kuhrman Jul 26 at 5:24

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.