Lee states that there will be many possible atlases that give rise to the same smooth structures on a manifold $M$; that is they determine the same collection of smooth functions on $M$.

Unfortunately, I don't see how different atlases will give rise to different collections of smooth functions in the first place.

For example, I'm struggling to think of two atlases on $\mathbb{R}$ which would not give the same collection of smooth functions; namely the ones in the sense of ordinary calculus.

Could someone provide me with an example of a manifold and two atlases on this manifold such that the collection of smooth functions determined by the atlases are not equivalent?


1 Answer 1


Let $\{(U_\alpha,\phi_\alpha)\}_{\alpha\in\mathcal{A}}$ and $\{(U_\beta,\phi_\beta)\}_{\beta\in\mathcal{B}}$ be two smooth (that is $\mathcal{C}^{\infty}$) atlases on a topological manifold $M$. By definition the $\mathcal{C}^{\infty}$-structures defined by these two atlases are equivalent if their union $\{(U_\alpha,\phi_\alpha)\}_{\alpha\in\mathcal{A}\cup\mathcal{B}}$ is also a smooth atlas, that is, the transition maps between the charts of different atlases are smooth. It can be shown, that this definition, is equivalent to saying that two $\mathcal{C}^{\infty}$-structures are equivalent if they determine the same set of smooth functions $f:M\rightarrow\mathbb{R}$.

Here is an example of two non-equivalent $\mathcal{C}^{\infty}$-structures:

Consider the topological manifold $\mathcal{N} = \mathbb{R}$ equipped with the $\mathcal{C}^{\infty}$-structure determined by the $\mathcal{C}^{\infty}$-atlas consisting of the single chart $(\mathbb{R}, Id)$.

Next consider $\mathbb{R}$ equipped with the $\mathcal{C}^{\infty}$-structure determined by the $\mathcal{C}^{\infty}$-atlas consisting of the single chart $(\mathbb{R}, \phi)$ where $\phi(x) = x^{3}$.

We can now show, that each one of the above $\mathcal{C}^{\infty}$-atlases (and thus: their $\mathcal{C}^{\infty}$-structures) are non-equivalent (because, for example, the change of coordinates: $[Id\circ\phi^{-1}](x)=x^{1/3}$ is not smoothly differentiable at $0$), therefore, the $\mathcal{C}^{\infty}$-structures determined by them are different.

  • $\begingroup$ Ok. So atlas 1 givesthe collection of all smooth functions on $\mathbb{R}$ as in calculus, while atlas 2 gives the collection of functions which are smooth when composed with $x^{\frac13}$, so for example $x^3$. $\endgroup$
    – user7090
    Sep 23, 2016 at 3:52
  • $\begingroup$ Yes, this is what I had in mind. $\endgroup$
    – KonKan
    Sep 23, 2016 at 4:14

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.