Corollary 9 in these notes by Ralph Cohen has grabbed my attention.

I do not undestand how to show that if we have a rank $k$ bundle which is stably isomorphic to the stable normal bundle then there is a virtual normal bundle of rank $k$.

This seems to boil down to proving the following thing:

Let $M^m$ a smooth manifold and suppose we are given $f:M^m\to BO(k)$. Let $n>k$ and let $g: M^m\to BO(n)$ be such that

$$TM\oplus g^*(EO(n)) \simeq \varepsilon^{n+m}.$$

In other words, $g$ is the classifying map for a virtual normal bundle of rank $n$ ($\varepsilon$ is the trivial bundle).

Suppose that $g$ is homotopic to $i\circ f$ where $i:BO(k)\to BO(n)$ is the obvious inclusion. Then we would like to show that

$$TM\oplus f^*(EO(k)) \simeq \varepsilon^{k+m}.$$

However the only thing I can conclude from the homotopy $g\simeq i\circ f$ is that

$$TM\oplus f^*(EO(k))\oplus \varepsilon^{n-k} \simeq \varepsilon^{n+m},$$

which is not enough in general.

  • $\begingroup$ What's a "virtual normal bundle"? (it seems from your description of g that it's synonymous with "stable normal bundle"?) $\endgroup$
    – kiran
    Jun 22 at 15:51
  • 1
    $\begingroup$ @kiran Thank you for your comment, I borrowed this def. from Cohen's notes. In general given an immersion $f:M\to N$, a virtual normal bundle for $f$ is a bundle $\nu$ such that $TM\oplus \nu \simeq f^*N$. In the above setting we are dealing with immersions into Euclidean spaces so we can say that $\nu$ is virtual normal bundle of rank $k$ if $TM\oplus \nu\simeq \varepsilon^{n+k}$, this implies by Hirsch-Smale theory that $M$ has an immersion into $\mathbb R^{n+k}$ with normal bundle isomorphic to $\nu$. If $k$ is large, $\nu$ is isomorphic to the stable normal bundle, for small $k$ idk. $\endgroup$ Jun 22 at 16:00

1 Answer 1


This follows from obstruction theory; also see this answer.

If $E \to X$ is a rank $r$ real vector bundle over a CW complex $X$, then the obstructions to finding a nowhere-zero section lie in $H^i(X; \pi_{i-1}(S^{r-1}))$. In particular, if $r > \dim X$, then such a section exists so $E\cong E_0\oplus\varepsilon^1$ where $E_0$ has rank $r - 1$. However, $E_0$ may not be unique. For example, the bundle $\varepsilon^{n+1} \to S^n$ has rank $r = n+1 > n$ and decomposes as $\varepsilon^n\oplus\varepsilon^1$ and $TS^n\oplus\varepsilon^1$.

The obstructions to the uniqueness of a nowhere-zero section up to scale lie in $H^i(X; \pi_i(S^{r-1}))$. In particular, if $r > \dim X + 1$, then $E$ admits a unique nowhere-zero section up to scale, so $E \cong E_0\oplus\varepsilon^1$ where $E_0$ has rank $r - 1$ and is unique up to isomorphism. Note, in the example above, $r = \dim X + 1$. If the rank of $E_0$, namely $r - 1$, is still larger than $\dim X + 1$, then we can apply the same argument to split off a trivial line bundle with a unique complement. It follows that if $V\oplus\varepsilon^p \cong W\oplus\varepsilon^p$ and $\operatorname{rank} V = \operatorname{rank} W > \dim X$, then $V \cong W$. In particular, if $\operatorname{rank} V > \dim X$ is stably trivial, then it is in fact trivial.

Since $TM\oplus f^*(EO(k))$ is stably trivial and $TM\oplus f^*(EO(k))$ has rank $m + k > m = \dim M$, we see that $TM\oplus f^*(EO(k)) \cong \varepsilon^{m+k}$.


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.