Arc Length (Unit 8 Applications of Integration, AP Calculus BC)

What Arc Length Means for a Curve

Arc length is the total distance you would travel if you traced along a curve from one endpoint to another. For straight-line segments, distance is easy—you use the distance formula. For a curved path, the key idea is to approximate the curve by many tiny straight segments and add up their lengths.

This matters in calculus because it is a classic “applications of integration” problem: you use an integral to add up infinitely many tiny contributions. It also connects directly to the deeper geometric idea of measuring curves (a major reason derivatives and integrals were developed historically).

From “add up small segments” to a calculus formula

Imagine a smooth curve in the plane (no sharp corners) and suppose you want the length from a starting point to an ending point.

1. Approximate the curve with a polygonal path. Pick points along the curve and connect consecutive points with straight lines.
2. Compute each small straight length using the Pythagorean theorem. If one tiny segment changes in the horizontal direction by $\Delta x$ and vertically by $\Delta y$, then its length is approximately

$\sqrt{(\Delta x)^2 + (\Delta y)^2}$

3. Add them up. If you use many points, the polygonal path hugs the curve more closely.
4. Take a limit as the segments get shorter and more numerous. That limit becomes an integral.

The “smoothness” condition is important. A smooth planar curve typically means the curve is differentiable and its derivative behaves nicely (often continuous) on the interval. Smoothness ensures the small straight-segment approximation converges to a well-defined length. If there’s a corner or cusp, you usually have to split the curve into pieces where it is smooth.

The differential arc length idea

For a curve given as $y=f(x)$, over a tiny change $dx$, the vertical change is approximately

$dy \approx f'(x)\,dx$

A tiny piece of arc length $ds$ (think: “small length”) is then approximated by

$ds \approx \sqrt{(dx)^2 + (dy)^2}$

Factor out $dx$ (conceptually, not as a formal algebraic step yet): the length depends on how much the curve rises/falls compared to how much it runs. That is why the derivative appears in the final formula.

Why arc length integrals can be hard

Arc length formulas often produce integrals involving square roots like

$\int \sqrt{1+(f'(x))^2}\,dx$

These frequently do not have elementary antiderivatives. On AP Calculus BC, you are often asked to set up the correct integral, and sometimes to evaluate it when it simplifies nicely (or to approximate it numerically when allowed).

Exam Focus

• Typical question patterns:
  • “Set up an integral for the length of the curve $y=f(x)$ from $x=a$ to $x=b$.”
  • “Find the arc length of a curve given parametrically on $t\in[a,b]$.”
  • “Compute arc length for a polar curve over a given $\theta$ interval.”
• Common mistakes:
  • Treating arc length like area and using $\int f(x)\,dx$ instead of the arc length integrand.
  • Forgetting that arc length requires a square root of a sum of squares (coming from the Pythagorean theorem).
  • Not splitting the interval when the curve is only smooth on pieces.

Arc Length for a Graph $y=f(x)$ on $[a,b]$

When a curve is given explicitly as a function of $x$—that is, $y=f(x)$—you can measure its length from $x=a$ to $x=b$ by adding the lengths of tiny hypotenuses formed by small horizontal and vertical changes.

The arc length formula (Cartesian, function of $x$)

If $f$ is differentiable on $[a,b]$, the arc length $L$ is

$L=\int_a^b \sqrt{1+(f'(x))^2}\,dx$

Here’s what each part means:

• $L$ is the total length along the curve.
• $f'(x)$ measures slope (how fast $y$ changes compared to $x$).
• The expression $\sqrt{1+(f'(x))^2}$ is the “stretch factor” converting a tiny horizontal change $dx$ into a tiny along-the-curve length $ds$.

A useful way to interpret it: if the curve is nearly flat (small slope), then $f'(x)\approx 0$ and the integrand is near $\sqrt{1}=1$, so arc length is close to the horizontal distance $b-a$. If the curve is steep, the integrand becomes larger, reflecting a longer path.

When to use this formula

Use this formula when:

• The curve is naturally described as $y$ as a function of $x$.
• The interval is specified in $x$-values (from $x=a$ to $x=b$).

If the curve fails to be a function of $x$ (for instance, it doubles back horizontally), you should consider a parametric description or splitting into pieces.

Worked example 1 (evaluates nicely)

Find the length of the curve

$y=\frac{2}{3}x^{3/2}$

from $x=0$ to $x=1$.

Step 1: Differentiate.

$y'=\frac{2}{3}\cdot \frac{3}{2}x^{1/2}=x^{1/2}$

So

$1+(y')^2=1+x$

Step 2: Set up the arc length integral.

$L=\int_0^1 \sqrt{1+x}\,dx$

Step 3: Evaluate.

$\int \sqrt{1+x}\,dx=\int (1+x)^{1/2}\,dx=\frac{2}{3}(1+x)^{3/2}+C$

So

$L=\frac{2}{3}(1+x)^{3/2}\Big|_0^1=\frac{2}{3}\left(2^{3/2}-1\right)$

That is the exact length.

Worked example 2 (set up only, because it becomes difficult)

Set up (but do not attempt to evaluate in elementary form) the arc length of

$y=\sin(x^2)$

from $x=0$ to $x=1$.

Step 1: Differentiate.

$y'=\cos(x^2)\cdot 2x$

Step 2: Plug into the arc length formula.

$L=\int_0^1 \sqrt{1+\left(2x\cos(x^2)\right)^2}\,dx$

There is no obvious simplification to make this elementary. On many calculus assessments (including AP-style problems), the goal is to produce this correct setup.

What commonly goes wrong with $y=f(x)$ arc length

A frequent misconception is thinking that “length” should involve $\int \sqrt{1+f(x)^2}\,dx$. The function value $f(x)$ tells you height, not the local steepness. Arc length depends on how the curve is changing, so the derivative must appear.

Another common issue: students sometimes try to cancel the square root incorrectly. Remember the structure comes from distance:

$\text{distance}^2=(\text{horizontal})^2+(\text{vertical})^2$

You should expect a square root of a sum of squares.

Exam Focus

• Typical question patterns:
  • “Compute the arc length of $y=f(x)$ on $[a,b]$.” (Often chosen so the integral simplifies.)
  • “Set up the arc length integral for a given function.”
  • “Decide whether to split the interval if $f'(x)$ is not continuous everywhere.”
• Common mistakes:
  • Using $f(x)$ instead of $f'(x)$ inside the formula.
  • Forgetting the square root or forgetting the $+1$ in $\sqrt{1+(f'(x))^2}$.
  • Using the wrong bounds (for example, using $y$-values instead of $x$-values when the problem specifies an $x$ interval).

Arc Length for a Graph $x=g(y)$ on $[c,d]$

Sometimes a curve is more naturally expressed as $x$ as a function of $y$ (for example, sideways parabolas). The same geometric idea applies: you approximate with tiny straight segments, but now it’s convenient to use vertical steps $dy$.

The arc length formula (Cartesian, function of $y$)

If $x=g(y)$ is differentiable on $[c,d]$, the arc length is

$L=\int_c^d \sqrt{1+\left(g'(y)\right)^2}\,dy$

Here $g'(y)=\frac{dx}{dy}$ measures how rapidly the curve moves horizontally as $y$ changes. This formula is essentially the same as the $y=f(x)$ version, but with the roles of $x$ and $y$ swapped.

When this is the right tool

Use this when:

• The curve is given explicitly as $x$ in terms of $y$.
• The interval is given in $y$-values.

If you instead try to solve for $y$ as a function of $x$, you might create a function that is not single-valued (fails the vertical line test) or you might complicate the algebra unnecessarily.

Worked example (sideways parabola)

Find the length of the curve

$x=y^2$

from $y=0$ to $y=2$.

Step 1: Differentiate with respect to $y$.

$\frac{dx}{dy}=2y$

Step 2: Set up the arc length integral.

$L=\int_0^2 \sqrt{1+(2y)^2}\,dy=\int_0^2 \sqrt{1+4y^2}\,dy$

This integral can be evaluated using a hyperbolic substitution or a known formula, but on an AP-style exam you might be asked only for the setup unless it’s designed to simplify.

Common pitfalls

A typical error is mixing differentials: writing something like

$\int \sqrt{1+\left(\frac{dx}{dy}\right)^2}\,dx$

The variable of integration must match the derivative inside. If you use $\frac{dx}{dy}$, then you integrate with respect to $y$.

Exam Focus

• Typical question patterns:
  • “A curve is given by $x=g(y)$. Set up the arc length on $y\in[c,d]$.”
  • “Choose whether it’s easier to use $y=f(x)$ or $x=g(y)$ for a given curve.”
• Common mistakes:
  • Using $dx$ when the setup is in terms of $dy$.
  • Converting to $y=f(x)$ and accidentally losing part of the curve (for example, taking only the top branch).

Arc Length for Parametric Curves $x=x(t),\;y=y(t)$

Parametric equations describe a curve by giving both coordinates as functions of a parameter $t$. This is especially powerful when:

• the curve loops or fails the vertical line test,
• motion is involved (the parameter behaves like time),
• or the algebra is simpler parametrically.

Why the parametric arc length formula looks the way it does

Over a small change $\Delta t$, the changes in coordinates are approximately

$\Delta x \approx x'(t)\,\Delta t$

and

$\Delta y \approx y'(t)\,\Delta t$

The tiny straight-line distance traveled is approximately

$\sqrt{(\Delta x)^2+(\Delta y)^2}$

Substituting the approximations and taking a limit produces the integral.

The parametric arc length formula

If $x(t)$ and $y(t)$ are differentiable on $[a,b]$, then the arc length is

$L=\int_a^b \sqrt{\left(\frac{dx}{dt}\right)^2+\left(\frac{dy}{dt}\right)^2}\,dt$

Interpretation: the integrand is the speed along the curve (distance per unit $t$). This connects arc length to kinematics: if $t$ is time, then arc length is total distance traveled.

Worked example (simplifies cleanly)

Find the arc length of the parametric curve

$x=3t^2$

$y=2t^3$

for $t$ from $0$ to $1$.

Step 1: Differentiate each component.

$\frac{dx}{dt}=6t$

$\frac{dy}{dt}=6t^2$

Step 2: Plug into the formula.

$L=\int_0^1 \sqrt{(6t)^2+(6t^2)^2}\,dt=\int_0^1 \sqrt{36t^2+36t^4}\,dt$

Factor inside the square root:

$L=\int_0^1 \sqrt{36t^2(1+t^2)}\,dt=\int_0^1 6t\sqrt{1+t^2}\,dt$

(Here $t\ge 0$ on $[0,1]$, so $\sqrt{t^2}=t$.)

Step 3: Evaluate with substitution. Let

$u=1+t^2$

Then

$du=2t\,dt$

So

$L=6\int_0^1 t\sqrt{1+t^2}\,dt=6\cdot \frac{1}{2}\int_{u=1}^{u=2} u^{1/2}\,du=3\int_1^2 u^{1/2}\,du$

Compute:

$\int u^{1/2}\,du=\frac{2}{3}u^{3/2}+C$

Thus

$L=3\cdot \frac{2}{3}u^{3/2}\Big|_1^2=2\left(2^{3/2}-1\right)$

A note about reparameterization

Arc length depends on the geometric path, not on how fast you move along it. If you change the parameterization (for example, replace $t$ with $2t$ and adjust the interval), the integral may look different but the final length should be the same—assuming you trace the same portion of the curve once.

Common pitfalls with parametric arc length

• Forgetting to square derivatives: The formula uses a sum of squares, so negative derivatives do not create negative length.
• Using the wrong interval: The parameter bounds are given in $t$, not $x$ or $y$.
• Not checking for retracing: Some parameterizations trace the same curve segment more than once; then the arc length integral gives total distance traveled, which would count retraced parts multiple times.

Exam Focus

• Typical question patterns:
  • “Given $x(t)$ and $y(t)$, find the length on $t\in[a,b]$.”
  • “Set up the arc length integral for a parametric curve; evaluate if possible.”
  • “A particle follows a parametric path; interpret arc length as total distance traveled.”
• Common mistakes:
  • Plugging $x(t)$ and $y(t)$ into the $y=f(x)$ formula directly instead of using $\frac{dx}{dt}$ and $\frac{dy}{dt}$.
  • Dropping the square root too early or mishandling $\sqrt{t^2}$ (you must consider the sign of $t$ on the interval).

Arc Length for Polar Curves $r=r(\theta)$ (A Planar Curve in Polar Form)

Polar coordinates describe a planar curve using a radius $r$ from the origin and an angle $\theta$. Many curves with rotational or circular symmetry are simpler in polar form.

Where the polar arc length formula comes from

A small change in angle $d\theta$ creates two kinds of change:

• a change in radius (outward/inward) of size about $dr$,
• and a change due to “sweeping” around the circle, which contributes about $r\,d\theta$.

These two changes are perpendicular in the limit, so a tiny arc length element satisfies an approximate Pythagorean relationship:

$ds\approx \sqrt{(dr)^2+(r\,d\theta)^2}$

Dividing by $d\theta$ and taking limits leads to the polar formula.

The polar arc length formula

If $r(\theta)$ is differentiable on $[\alpha,\beta]$, then the arc length is

$L=\int_{\alpha}^{\beta} \sqrt{\left(r(\theta)\right)^2+\left(\frac{dr}{d\theta}\right)^2}\,d\theta$

This is a direct analogue of the parametric formula. In fact, you can justify it by converting to parametric form:

$x=r(\theta)\cos(\theta)$

$y=r(\theta)\sin(\theta)$

and using the parametric arc length formula.

Worked example (nice simplification)

Find the arc length of the polar curve

$r=2\cos(\theta)$

for $\theta$ from $0$ to $\frac{\pi}{2}$.

Step 1: Differentiate.

$\frac{dr}{d\theta}=-2\sin(\theta)$

Step 2: Plug into the polar arc length formula.

$L=\int_0^{\pi/2} \sqrt{(2\cos(\theta))^2+(-2\sin(\theta))^2}\,d\theta$

Simplify inside the root:

$L=\int_0^{\pi/2} \sqrt{4\cos^2(\theta)+4\sin^2(\theta)}\,d\theta=\int_0^{\pi/2} \sqrt{4}\,d\theta=\int_0^{\pi/2} 2\,d\theta$

Step 3: Integrate.

$L=2\theta\Big|_0^{\pi/2}=\pi$

Common pitfalls with polar arc length

• Mixing up the formula with polar area (which is $\frac{1}{2}\int r^2\,d\theta$). Arc length uses $\sqrt{r^2+(dr/d\theta)^2}$, not $r^2$.
• Using the wrong bounds in $\theta$, especially if the curve has symmetry. If you use symmetry to shorten work, you must be sure the curve segment is traced exactly once.

Exam Focus

• Typical question patterns:
  • “Find the length of $r=r(\theta)$ on a specified $\theta$ interval.”
  • “Set up the integral for polar arc length; evaluate if it simplifies.”
• Common mistakes:
  • Confusing polar arc length with polar area.
  • Forgetting to compute $dr/d\theta$ or squaring it incorrectly.

Choosing the Right Arc Length Setup (and Knowing When to Split)

Even though there are multiple formulas, they all come from the same geometric fact: small pieces of a smooth curve behave like hypotenuses of tiny right triangles.

Notation and formula reference (same idea, different forms)

When you must split the interval

Arc length is additive: if a curve is smooth on $[a,c]$ and $[c,b]$, then

$L=\int_a^c (\cdots) + \int_c^b (\cdots)$

You typically split when:

• The derivative is undefined at a point (a cusp, corner, or vertical tangent in a Cartesian setup).
• A parametric curve changes direction in a way that makes algebraic simplification depend on sign (for example, when simplifying $\sqrt{t^2}$).
• A polar curve crosses the origin or is traced multiple times, and you need to ensure you are measuring exactly the intended portion.

Reality check: exact vs. numerical answers

In many authentic applications (engineering, physics, computer graphics), arc length is computed numerically because the integrals are hard. In an AP context, you should be comfortable with both:

• Exact evaluation when the integrand simplifies (often by algebra, substitution, or a trig identity).
• Correct setup when it does not.

Exam Focus

• Typical question patterns:
  • “Here is a curve (possibly in an unfamiliar form). Choose the correct arc length formula and set up the integral.”
  • “Compute the length and show correct work; simplification is often the main hurdle.”
• Common mistakes:
  • Using bounds from the wrong variable (for example, using $x$-bounds in a $d\theta$ integral).
  • Not recognizing when a curve is traced twice (leading to an arc length that is double what the question intends).
  • Failing to split when simplifications require sign information (especially with $\sqrt{t^2}$ behavior across negative and positive $t$).