Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. 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

Let $\alpha:[a,b] \rightarrow \mathbb{R}^n$ be a path and let $\alpha^{-1}$ denote the inverse path given by $\alpha^{-1}(t)=\alpha(a + b - t)$ Also, let $F$ be a vector field on $\mathbb{R}^n$ and assume appropriate hyphotheses regarding smoothness, domains, etc. I am trying to show that

$$ \int\limits_{\alpha^{-1}} F \cdot d(\alpha^{-1}) = -\int\limits_{\alpha} F \cdot d\alpha $$

This is an intuitive result but, technically, I can't seem to manage to get the signs right and I'm sure I'm making a silly mistake. Here are my computations:

$$ \int\limits_{\alpha^{-1}} F \cdot d(\alpha^{-1}) = $$

$$ =\int\limits_{a}^{b}F(\alpha^{-1}(t)) \cdot [\alpha^{-1}(t)]^\prime dt $$

$$ =\int\limits_{a}^{b}F(\alpha(a + b -t)) \cdot \alpha^{\prime}(a + b - t)dt $$

Now, let $u = a + b - t$. Then, $t = a \implies u = b$ and $t = b \implies u = a$. Furthermore, $dt = -du$. Therefore, we can rewrite the last integral as

$$ - \int\limits_{b}^{a}F( \alpha(u)) \cdot \alpha^{\prime}(u)du $$

$$ = \int\limits_{a}^{b}F( \alpha(u)) \cdot \alpha^{\prime}(u)du $$

$$ = \int\limits_{\alpha} F \cdot d\alpha $$

Which is, obviously, not what I'm trying to show. There are two sign changes here which cancel one another out. The first is due to the fact that $dt = -du$ and the second because I am reversing the order the integration limits so that the integral fits the definition of the path integral.

So, my question is, what step is in error?

share|cite|improve this question
up vote 3 down vote accepted

The mistake is in the derivative of $\alpha^{-1}$. That is from step 2 to 3 in your reasoning. It should be $-\alpha'(a+b-t)$.

share|cite|improve this answer
But, from step 2 to step 3 I'm not actually evaluating the derivative; I'm just replacing $\alpha^{-1}$ with $\alpha(a + b - t)$. That is, $\frac{d}{dt}( \alpha^{-1}(t) ) = \frac{d}{dt}(\alpha(a + b - t))$ – ItsNotObvious Jun 18 '11 at 22:27
Yes, but you have to apply the chain rule. The derivative of $\alpha(a+b-t)$ is that of $\alpha$ evaluated in $a+b-t$ times the derivative of $a+b-t$ with respect to $t$. – Raskolnikov Jun 19 '11 at 8:43
OK, I think I understand the situation after thinking about it a little more. Basically, the derivative of $\alpha$ must actually taken in order for the following substitution step to actually make sense. Thanks. – ItsNotObvious Jun 19 '11 at 13:49

I think you forgot to apply the chain rule when taking the derivative of $\alpha(a+b-t)$. There should be an additional factor $-1$ from the $-t$ term.

share|cite|improve this answer

If I remember well , using c for the curve, you can refer to the original integrand as f(c(a+t(b-a)))c'(a+t(b-a)), since this parametrization agrees with the standard ordering of the real line, i.e., the parametrization a+t(b-a) will output the points in the order you want them. Then the reverse orientation is the reverse ordering-- b-t(a-b)--but the image f(c(b-t(a-b)) is still the same as the original; f(x) is the same, only seen from a different direction. Note that your choices of parametrization for the curves c and $c^-1$ do not cancel out to the identity (giving you the starting point). Then, by the chain rule , c'(b-t(b-a)) will give you the negative sign.

share|cite|improve this answer

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.