Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. Join them; it only takes a minute:

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

A continuous function $\alpha: [a,b] \to \mathbb{R}^k$ is called a curve. For each partition $P = \{t_0<t_1<....<t_n=b\}$, define $l(\alpha, P) = \sum_{i=1}^n \left|\alpha(t_i) - \alpha(t_{i-1})\right|$.

Let $l(\alpha) = \sup\{l(\alpha, P): P \text{ partitions } [a,b]\}$

If $l(\alpha) < \infty$, then $\alpha$ is called rectifiable.

a) Let $c < a < b < d.$ Suppose $\alpha'$ is continuous on $(c,d)$ for some curve $\alpha: [c,d] \to \mathbb{R}^k$. Show that the restriction $\alpha|_{[a,b]}$ of $\alpha$ to $[a,b]$ is rectifiable with $l(\alpha|_{[a,b]}) = \int_a^b \left|\alpha'(t)\right|\, dt$.

b) Let $c < a < b < d.$ Suppose $f: (c,d) \to \mathbb{R}$ is continuously differentiable. Let $\alpha: [a,b] \to \mathbb{R}^2$ be given by $\alpha(t) = (t,f(t))$. Find $l(\alpha)$ in terms of $f$.

Here's what I have so far.

a) we can use the Cauchy-Schwarz inequality, unsure how to implement it, and not sure what else we can invoke b) If $f$ is continuously differentiable, can I talk about uniform convergence, will it help? I am stuck on this one.

share|cite|improve this question
+1 for explaining what you have tried. – Jesse Madnick Apr 26 '11 at 8:27
For b), you just need to apply the formula found in a) in your specific case. – Plop Apr 26 '11 at 11:05
up vote 4 down vote accepted

I'm expanding my original answer; so its no longer a hint anymore.

The inequality $l(\alpha)\leq\int_a^b|\alpha'(t)|dt\ $ is easy: One has $$|\alpha(t_i)-\alpha(t_{i-1})| = \Bigl |\int_{t_{i-1}}^{t_i} \alpha'(t)dt \Bigr| \leq \int_{t_{i-1}}^{t_i} |\alpha'(t)|dt \ ,$$ and by summation we obtain $$l(\alpha,P)\leq \int_a^b |\alpha'(t)|dt\ .$$ Since this is true for all partitions $P$ the inequality follows.

For the other direction a trick is required. Let an $\epsilon>0$ be given and assume that the partition $P$ is fine enough to guarantee $|\alpha'(t)-\alpha'(t')|<\epsilon$ within each subinterval. Then for each $i$ one has $$\int_{t_{i-1}}^{t_i}|\alpha'(t)|dt=|\alpha'(\tau)|(t_i-t_{i-1}) =\Bigl|\int_{t_{i-1}}^{t_i}\alpha'(\tau) dt\Bigr| \leq \Bigl|\int_{t_{i-1}}^{t_i}\alpha'(t) dt\Bigr| +\epsilon(t_i-t_{i-1})\ ,$$ where $\tau$ is a certain ${\it fixed}$ point in the interval $[t_{i-1},t_i]$. Summing over $i$ we get $$\int_a^b|\alpha'(t)|dt\leq l(\alpha, P)+\epsilon(b-a)\leq l(\alpha)+\epsilon (b-a)\ .$$ As $\epsilon>0$ was arbitrary we conclude $\int_a^b|\alpha'(t)|dt\leq l(\alpha)\ .$

share|cite|improve this answer
I'm not sure how to write alpha(tsubi) - alpha(tsubi-1) in terms of alpha prime. Can you explain, please? – mary Apr 27 '11 at 4:06
Can you please expand upon further for the problem? I am lost – mary Apr 28 '11 at 6:08

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.