Rectification and geometric applications

Rectification is the classical name for finding the length of a curve using integral calculus — literally "straightening" a bent path into an equivalent straight-line length. It sits inside the "Integral calculus and multiple integrals" unit of MA10002, right after the analytical machinery (reduction formulae, Beta and Gamma functions, Leibnitz's rule) and just before you start using the definite integral to measure genuinely geometric things: not just area under a curve, but the length of the curve itself, and later the areas and volumes swept out when curves revolve. The idea is the same one behind every application of the definite integral — chop the curve into tiny pieces, use Pythagoras on each piece, and sum (integrate) the results. What changes from problem to problem is only which form the curve is given in — Cartesian $y=f(x)$ , parametric $x(t),y(t)$ , or polar $r(\theta)$ — because each form leads to a differently dressed version of the same square-root integrand.

The core insight

Arc length is the integral of the infinitesimal Pythagorean distance $\sqrt{(dx)^2+(dy)^2}$ along the curve, rewritten in whichever variable — $x$ , $t$ , or $\theta$ — the curve is naturally parametrised by.

The core idea

Cartesian form. For a smooth curve $y=f(x)$ on $[a,b]$ , a tiny piece of the curve over a step $dx$ has horizontal run $dx$ and vertical rise $dy = y'\,dx$ , so by Pythagoras its length is $\sqrt{(dx)^2+(dy)^2} = \sqrt{1+(y')^2}\,dx$ . Summing every such piece gives the arc length

s = \int_a^b \sqrt{1+\left(\frac{dy}{dx}\right)^2}\,dx,

where: $s$ is the arc length, and $y' = dy/dx$ is the slope of the curve at each point.

Parametric form. When the curve is handed to you as $x=x(t)$ , $y=y(t)$ for $t\in[t_1,t_2]$ , do not fight to eliminate $t$ and force it back into $y=f(x)$ — differentiate both coordinates with respect to $t$ instead:

s = \int_{t_1}^{t_2} \sqrt{\left(\frac{dx}{dt}\right)^2+\left(\frac{dy}{dt}\right)^2}\,dt,

where: $dx/dt$ and $dy/dt$ are the component speeds along each axis; the integrand is again just $\sqrt{(dx)^2+(dy)^2}$ , now factored out by $dt$ .

Polar form. For $r=r(\theta)$ on $[\theta_1,\theta_2]$ ,

s = \int_{\theta_1}^{\theta_2} \sqrt{r^2+\left(\frac{dr}{d\theta}\right)^2}\,d\theta,

where: $r$ is the radius at angle $\theta$ and $dr/d\theta$ its rate of change; this follows from the parametric formula via $x=r\cos\theta$ , $y=r\sin\theta$ , differentiated with the product rule. The one decision that matters before you touch an integral sign is which of these three to use, and that is dictated purely by how the curve is presented: $y=f(x)$ calls for the Cartesian form, a pair $(x(t),y(t))$ stays parametric, and a spiral- or cardioid-type curve given as $r(\theta)$ is far easier to rectify directly in polar form than after converting to Cartesian coordinates. Trying to force a parametric or polar curve back into $y=f(x)$ before integrating is almost always the slower path — sometimes the elimination is algebraically painful, and sometimes it is not possible at all, so recognising the given form early saves real time in an exam. All three formulas are, underneath, the same statement: arc length is a definite integral of the local speed of tracing the curve, whether that tracing is with respect to $x$ , an abstract parameter $t$ , or the polar angle $\theta$ . This is exactly the same "sum of infinitesimal pieces" logic that gives area under a curve — only here the pieces being summed are lengths of tiny line segments rather than thin rectangles of area, so the resulting integrand always carries a square root rather than a bare function value.

Figure

Three tiny curve segments each shown as a hypotenuse of a right triangle with legs dx and dy labelled, one drawn under a Cartesian curve, one under a parametric curve, one under a polar curve, illustrating that all three arc length formulas come from the same Pythagorean step

Derivation AThe parametric arc-length formula and the cycloid

The goal is to derive the general parametric arc-length formula from the Pythagorean step and then apply it to find the length of one complete arch of the cycloid $x=a(\theta-\sin\theta)$ , $y=a(1-\cos\theta)$ for $\theta\in[0,2\pi]$ , where $a>0$ is a constant, reducing a square root that looks intractable at first glance into a single clean trigonometric term using a half-angle identity.

Step 1. Start from the infinitesimal right-triangle length $ds=\sqrt{(dx)^2+(dy)^2}$ used for any smooth curve. If $x$ and $y$ are both functions of a single parameter $t$ , factor $dt^2$ out from under the root: $ds=\sqrt{(dx/dt)^2+(dy/dt)^2}\,dt$ . Summing (integrating) these pieces from $t_1$ to $t_2$ gives $s=\int_{t_1}^{t_2}\sqrt{(dx/dt)^2+(dy/dt)^2}\,dt$ , the parametric formula.

Step 2. Apply it with $\theta$ as the parameter for the cycloid. Differentiate: $dx/d\theta = a(1-\cos\theta)$ and $dy/d\theta = a\sin\theta$ . Squaring and adding, $(dx/d\theta)^2+(dy/d\theta)^2 = a^2(1-\cos\theta)^2+a^2\sin^2\theta = a^2\left[1-2\cos\theta+\cos^2\theta+\sin^2\theta\right] = a^2[2-2\cos\theta]$ , using $\cos^2\theta+\sin^2\theta=1$ .

Step 3. Simplify the square root with the half-angle identity $1-\cos\theta=2\sin^2(\theta/2)$ : $a^2\cdot2(1-\cos\theta)=4a^2\sin^2(\theta/2)$ , so the integrand becomes $\sqrt{4a^2\sin^2(\theta/2)}=2a\left|\sin(\theta/2)\right|=2a\sin(\theta/2)$ for $\theta\in[0,2\pi]$ , since $\sin(\theta/2)\ge0$ on that whole range.

Step 4. Integrate: $s=\int_0^{2\pi}2a\sin(\theta/2)\,d\theta = 2a\left[-2\cos(\theta/2)\right]_0^{2\pi} = -4a\left[\cos\pi-\cos0\right] = -4a(-1-1) = 8a$ metres. One complete arch of the cycloid is exactly eight times the constant $a$ — a clean, memorable result that also serves as a check on the algebra. Verified: $(dx/d\theta)^2+(dy/d\theta)^2=4a^2\sin^2(\theta/2)$ , giving arc length $8a$ .

Worked example

Problem. Find the length of the curve $y=\dfrac{2}{3}x^{3/2}$ from $x=0$ to $x=3$ (all lengths in metres).

Step 1. Differentiate: $\dfrac{dy}{dx}=x^{1/2}$ , so $1+(y')^2 = 1+x$ .

Step 2. Set up the Cartesian arc-length integral: $s=\int_0^3\sqrt{1+x}\,dx$ .

Step 3. Substitute $u=1+x$ , $du=dx$ ; the limits become $u=1$ (at $x=0$ ) and $u=4$ (at $x=3$ ):

s=\int_1^4\sqrt{u}\,du=\left[\frac{2}{3}u^{3/2}\right]_1^4=\frac{2}{3}\left(4^{3/2}-1^{3/2}\right)=\frac{2}{3}(8-1)=\frac{14}{3}\text{ m}.

Answer: $s = \dfrac{14}{3}\text{ m} \approx 4.667\text{ m}$ .

Competitive stretch

The same half-angle trick from the derivation above also rectifies the cardioid $r=a(1+\cos\theta)$ , $\theta\in[0,2\pi]$ . Here $dr/d\theta=-a\sin\theta$ , so $r^2+(dr/d\theta)^2=a^2(1+\cos\theta)^2+a^2\sin^2\theta=a^2[2+2\cos\theta]=4a^2\cos^2(\theta/2)$ using $1+\cos\theta=2\cos^2(\theta/2)$ . Since $\cos(\theta/2)$ changes sign over $[0,2\pi]$ , compute the upper half ( $\theta\in[0,\pi]$ , where $\cos(\theta/2)\ge0$ ) and double it: $s=2\int_0^\pi 2a\cos(\theta/2)\,d\theta=4a\left[2\sin(\theta/2)\right]_0^\pi=8a\left[\sin(\pi/2)-\sin0\right]=8a$ metres — the same $8a$ as the cycloid, a useful cross-check that the absolute-value handling was done correctly.

Common traps

Forgetting the $1+$ inside the square root in the Cartesian formula — writing $s=\int y'\,dx$ instead of $s=\int\sqrt{1+(y')^2}\,dx$ .
Mixing up the parametric and Cartesian formulas — the parametric integral runs over $t$ (or $\theta$ ), with limits that are parameter values, not $x$ -values.
Dropping the absolute value when simplifying $\sqrt{\sin^2(\theta/2)}$ or $\sqrt{\cos^2(\theta/2)}$ -type expressions — always check the sign of the quantity inside the square root before removing it, and split the range if the sign changes, as in the cardioid.
In the polar formula, forgetting the extra $r^2$ term and using $(dr/d\theta)^2$ alone — both terms are needed under the square root.
Confusing rectification (arc length) with quadrature (area) — the two use structurally different integrands even though both are definite integrals.

Ask the tutor: if a half-angle simplification in an arc-length integrand is not falling into place, ask the AI tutor to walk through more trig-identity tricks used in rectification problems.

Why does the arc-length formula always have a square root in it?

Because it comes straight from the Pythagorean theorem. Each tiny piece of curve is treated as the hypotenuse of a right triangle with legs $dx$ and $dy$ (or their equivalents in $t$ or $\theta$ ), so its length is $\sqrt{(dx)^2+(dy)^2}$ . Integrating sums up all these tiny hypotenuses. The square root is not a stylistic choice — it is baked into the geometry of distance itself, which is why every version of the formula, Cartesian, parametric or polar, keeps that same square-root shape.

How do I decide which of the three arc-length formulas to use for a given problem?

Look at how the curve is written down, not at what it looks like. If you are handed $y=f(x)$ directly, use the Cartesian formula — converting it to another form only adds work. If the curve comes as a pair $x(t)$ , $y(t)$ , stay parametric even if you could in principle eliminate $t$ , because that elimination is often messy or impossible. If the curve is described as $r(\theta)$ , such as a spiral or a cardioid, use the polar formula directly rather than converting to Cartesian coordinates first, which usually makes the algebra far worse.

Why do the cycloid and the cardioid both come out to exactly 8a?

It is a coincidence of these two specific curves, not a general law — most curves give arc lengths with no such tidy closed form. Both problems happen to reduce, via the identities $1-\cos\theta=2\sin^2(\theta/2)$ and $1+\cos\theta=2\cos^2(\theta/2)$ , to an integrand that is a single trig function of $\theta/2$ scaled by $2a$ , and integrating that over the relevant range both times lands on $8a$ . It is a nice memorable check, useful for catching algebra slips, but do not expect it to generalise to other curves built from $a$ and $\theta$ .

Rectification and geometric applications

The core idea

Worked example

Competitive stretch

Common traps

Before this

Where this leads