I'm working on the following problem from Guillemin and Pollack's Differential Topology:

Let $c : \left[a, b\right] \rightarrow X$ be a smooth curve, and let $f: \left[a_1, b_1\right] \rightarrow \left[a, b\right]$ be a smooth map with $f(a_1) = a$ and $f(b_1) = b$. Show that the integrals

$$ \int _a ^b c^* \omega \text{ and } \int _{a_1} ^ {b_1} (c \circ f)^* \omega $$ are the same.

While it doesn't say so explicitly, I'm assuming $\omega$ is a compactly supported smooth $k$-form on the $k$-dimensional manifold $X$.

I know that if $c$ and $f$ are both orientation-preserving diffeomorphisms, then using a theorem in Guillemin and Pollack's book

$$ \int _{a_1} ^{b_1} (c \circ f)^*\omega = \int _{a_1} ^{b_1} f^* c^* \omega = \int _a ^b c^* \omega = \int _c \omega$$

However, I don't really see any reason to believe $c$ and $f$ are both orientation-preserving diffeomorphisms, so I'm not sure how to proceed. I know $c$ is a smooth map onto its image, but I can't see why it should be one-to-one. Can anyone provide some advice, please? In case it's relevant, this is Exercise 4 from Section 4.4 of Guillemin and Pollack's book, and this problem is for my own self study, not from a graded assignment.


$\omega$ should be a smooth $1$-form as you want to integrate this along a curve $c$. $\omega$ doesn't have to be of compact support, as $c$ is compact anyway.

Indeed, after the pullback $c:[a, b]\to X$, the calculations are done on $[a,b]$. Write $c^* \omega = g(x) dx$ on $[a,b]$. Then $$(c\circ f)^* \omega = f^* c^* \omega = f^* (g(x) dx) = g(f(x)) f'(x) dx.$$ So you are asking why $$\int_{a_1}^{b_1} g(f(x)) f'(x) dx = \int_a^b g(x) dx.$$ But this is true as a rule of substiution, which folllows from Chain rule. From the link you can also see how $f(a_1) = a$ and $f(b_1) = b$ are used.

  • $\begingroup$ Thank you for your answer. I just have one quick question. In my textbook it seems like in the definition of the integral of a differential form over a manifold it requires that the form is compactly supported. Why doesn't $\omega$ have to be of compact support here? $\endgroup$ – Gecko Nov 23 '15 at 3:37
  • 1
    $\begingroup$ If your manifold is $n$-dimensional and $\omega$ is a $n$-form, then you require that $\omega$ is of compact support. Now you are integrating $c^*\omega$ over $[a,b]$, so $c^*\omega$ has compact support on $[a,b]$ anyway. @Gecko $\endgroup$ – user99914 Nov 23 '15 at 3:54

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.