Skip to content
Marlbridge

AP® Calculus BC · Unit 9 · Topic 9.2

Second Derivatives of Parametric Equations: Study Guide (Calculus BC 9.2)

How to find d²y/dx² for a parametric curve by differentiating dy/dx with respect to t and dividing by dx/dt, and how to use it for concavity and tangent-line estimates.

BC-only material. Parametric equations are part of Calculus BC only. Calculus AB students do not need this topic.

Before you start: prerequisites

You need these earlier ideas. If any is shaky, revisit it from the Calculus BC roadmap first.

Prerequisite Topic What you use it for here
dy/dx = (dy/dt) ÷ (dx/dt) 9.1 The first step, and the same idea used again
Quotient rule and chain rule 2.9, 3.1 dy/dx is often a quotient of functions of t
Second derivatives and concavity 3.6, 5.6 Reading the sign of d²y/dx²
Tangent-line approximation 4.6 Deciding if an estimate is too big or too small

Notation on this page: x′(t) = dx/dt, y′(t) = dy/dt, and x″(t), y″(t) are the second derivatives with respect to t. Angles are in radians. You should already be confident with Topic 9.1.

The tempting shortcut that fails

For y = f(x), the second derivative d²y/dx² measures how the slope changes as x changes. Its sign gives the concavity.

For a parametric curve, it is tempting to differentiate x and y twice with respect to t and divide: d²y/dt² ÷ d²x/dt². This is wrong. Here is a quick test.

Take x = 2t, y = t². Eliminating t gives y = x²/4, so the true second derivative is d²y/dx² = ½ everywhere. But d²y/dt² = 2 and d²x/dt² = 0, so the shortcut gives 2 ÷ 0, which is undefined. The shortcut fails even on a simple parabola.

The reason: d²y/dx² means “the derivative with respect to x of dy/dx”. Nothing in the shortcut takes a derivative with respect to x.

Deriving the correct method

Call the slope m = dy/dx. In a parametric problem, m is a function of t. We want dm/dx, the rate at which the slope changes as x changes.

This is exactly the problem from Topic 9.1, with m in place of y. The chain rule gives dm/dt = (dm/dx)(dx/dt), so, provided dx/dt ≠ 0,

d²y/dx² = dm/dx = [d/dt (dy/dx)] ÷ (dx/dt)

Check with x = 2t, y = t²: dy/dx = 2t/2 = t. Then d/dt (t) = 1, and 1 ÷ (dx/dt) = 1 ÷ 2 = ½. ✓ This matches y = x²/4.

The method in four steps.

  1. Find dx/dt and dy/dt.
  2. Find dy/dx = (dy/dt) ÷ (dx/dt), and simplify it as a function of t.
  3. Differentiate dy/dx with respect to t.
  4. Divide the result by dx/dt.

Simplifying in step 2 saves a lot of quotient-rule work in step 3.

Background: a single formula. If you apply the quotient rule to (y′/x′) and divide by x′, you get d²y/dx² = (x′y″ − y′x″) ÷ (x′)³. It gives the same answer and can be used as a check. You do not need to memorise it.

What the sign tells you

Once you have d²y/dx², you read it exactly as for y = f(x):

Sign of d²y/dx² Shape near the point Tangent line nearby Tangent-line estimate
Positive Concave up Lies below the curve Underestimate
Negative Concave down Lies above the curve Overestimate

Watch the sign of dx/dt. In step 4 you divide by dx/dt. If dx/dt < 0 (the point is moving left), dividing flips the sign. So d/dt (dy/dx) > 0 does not always mean concave up. Worked example 2 shows this happening.

Worked example 1: a second derivative at a point

Question. A curve is given by x = 4 sin t, y = 2 cos 2t. Without a calculator, find d²y/dx² at t = π/6 and state the concavity there.

  1. First derivatives. dx/dt = 4 cos t. dy/dt = −4 sin 2t (chain rule: d/dt [cos 2t] = −2 sin 2t).
  2. dy/dx, simplified. Use sin 2t = 2 sin t cos t: dy/dx = (−4 sin 2t)/(4 cos t) = (−8 sin t cos t)/(4 cos t) = −2 sin t, for cos t ≠ 0.
  3. Differentiate with respect to t. d/dt (−2 sin t) = −2 cos t.
  4. Divide by dx/dt. d²y/dx² = (−2 cos t)/(4 cos t) = −½, for cos t ≠ 0.
  5. At t = π/6. The point is (4 · ½, 2 cos(π/3)) = (2, 1). The slope is −2 sin(π/6) = −1, and d²y/dx² = −½ < 0, so the curve is concave down there.

Check by eliminating t. cos 2t = 1 − 2 sin²t, so y = 2 − 4 sin²t = 2 − x²/4. Then dy/dx = −x/2 = −2 sin t ✓ and d²y/dx² = −½ ✓. The curve is part of a downward parabola.

The wrong shortcut here. d²y/dt² = −8 cos 2t = −4 and d²x/dt² = −4 sin t = −2 at t = π/6, giving (−4) ÷ (−2) = 2. That has the wrong sign and the wrong size, and would wrongly say “concave up”.

Worked example 2: concavity on intervals, and a tangent-line estimate

Question. A curve is given by x = t² + 2t and y = t³ − 3t (Figure 1). (a) Find dy/dx and d²y/dx² in terms of t. (b) For which values of t is the curve concave up? Concave down? (c) Use the tangent line at t = 0 to estimate y when x = 0.2 on the same part of the curve. Is the estimate too big or too small?

The curve x = t² + 2t, y = t³ − 3t, showing a concave-down part and a concave-up part Graph on x and y axes, x from −1.5 to 9.5 and y from −7 to 4.5. Both parts of the curve meet at a sharp point (a cusp) at (−1, 2) when t = −1. The part for t less than −1 comes in from the lower right near (0.96, −6.6), passes through (0, −2), and rises steeply to the cusp, bending so that it is concave down; an arrow shows it moving up and to the left. The part for t greater than −1 leaves the cusp going down and to the right, passes through the origin at t = 0, reaches a lowest point (3, −2) at t = 1, and then rises to (8, 2) and beyond, bending so that it is concave up; an arrow shows it moving to the right. A dashed tangent line through the origin with slope −3/2 lies just below the concave-up part. 2468 42−2−4−6 x y cusp at t = −1, point (−1, 2) t = 0 t = 1 tangent at t = 0 solid part, t > −1: concave up, d²y/dx² > 0 dashed part, t < −1: concave down, d²y/dx² < 0
Figure 1. The curve x = t² + 2t, y = t³ − 3t for −2.4 ≤ t ≤ 2.2. The part for t < −1 (long dashes) is concave down and moves up and left towards the cusp at (−1, 2); the part for t > −1 (solid) is concave up and moves to the right. The short-dashed line is the tangent at t = 0, which lies below the concave-up part.

(a) First and second derivatives.

  1. dx/dt = 2t + 2 = 2(t + 1). dy/dt = 3t² − 3 = 3(t − 1)(t + 1).
  2. dy/dx = 3(t − 1)(t + 1) ÷ 2(t + 1) = 3(t − 1)/2, for t ≠ −1.
  3. d/dt (dy/dx) = 3/2. This is positive for every t.
  4. d²y/dx² = (3/2) ÷ 2(t + 1) = 3 / (4(t + 1)), for t ≠ −1.

(b) Concavity.

  • For t > −1: t + 1 > 0, so d²y/dx² > 0 and the curve is concave up.
  • For t < −1: t + 1 < 0, so d²y/dx² < 0 and the curve is concave down.
  • At t = −1, dx/dt = 0 and dy/dt = 0, so dy/dx and d²y/dx² are not defined there. Figure 1 shows a sharp point (a cusp) at (−1, 2).

Notice that d/dt (dy/dx) = 3/2 is positive for all t, yet the curve is concave down for t < −1. The difference comes from dividing by dx/dt, which is negative when t < −1 (the point is moving left).

(c) Tangent-line estimate.

  1. At t = 0: the point is (0, 0), the slope is 3(0 − 1)/2 = −3/2, and d²y/dx² = 3/4.
  2. Tangent line: y = −1.5x. At x = 0.2 it gives y ≈ −0.3.
  3. d²y/dx² = 3/4 > 0, so the curve is concave up near t = 0. The tangent line lies below the curve, so the estimate −0.3 is an underestimate.

Checks.

  • Actual value: x = 0.2 gives t² + 2t − 0.2 = 0, so t = −1 + √1.2 ≈ 0.0954 on this part of the curve. Then y ≈ −0.2855, which is greater than −0.3. ✓ Underestimate.
  • Why “same part” matters: x = 0.2 also occurs at t = −1 − √1.2 ≈ −2.095 on the concave-down part, where y ≈ −2.91. The tangent line at t = 0 says nothing about that point.
  • Optional formula: (x′y″ − y′x″)/(x′)³ = [2(t + 1) · 6t − 3(t² − 1) · 2] / [8(t + 1)³] = 6(t + 1)² / [8(t + 1)³] = 3/(4(t + 1)). ✓

Worked example 3: when you are given only dx/dt and dy/dt

Many questions do not give x(t) and y(t) at all. They give only the two rates. That is enough, because the method never uses x or y themselves.

Question. A curve has dx/dt = t² − 4 and dy/dt = 2t. Without a calculator: (a) find d²y/dx² in terms of t; (b) find d²y/dx² at t = 1 and state the concavity there; (c) find the intervals of t on which the curve is concave up and concave down.

(a) Follow the four steps.

  1. dx/dt = t² − 4 and dy/dt = 2t are given.
  2. dy/dx = 2t/(t² − 4), for t ≠ ±2. Nothing cancels, so you need the quotient rule next.
  3. d/dt (dy/dx) = [2(t² − 4) − 2t · 2t] ÷ (t² − 4)² = (−2t² − 8)/(t² − 4)² = −2(t² + 4)/(t² − 4)².
  4. Divide by dx/dt = t² − 4: d²y/dx² = −2(t² + 4)/(t² − 4)³, for t ≠ ±2.

(b) At t = 1. dx/dt = −3 and d/dt (dy/dx) = −2(5)/9 = −10/9. So d²y/dx² = (−10/9) ÷ (−3) = 10/27 > 0, and the curve is concave up there. Notice the two negatives: dividing by a negative dx/dt flipped the sign.

(c) Sign analysis. The numerator −2(t² + 4) is always negative. So the sign of d²y/dx² is the opposite of the sign of (t² − 4)³, which has the same sign as t² − 4.

  • For −2 < t < 2: t² − 4 < 0, so d²y/dx² > 0 and the curve is concave up.
  • For t < −2 or t > 2: t² − 4 > 0, so d²y/dx² < 0 and the curve is concave down.

Checks. At t = 0, the formula gives −8/(−64) = 1/8 > 0 ✓, and at t = 3 it gives −26/125 < 0 ✓. At t = ±2, dx/dt = 0 but dy/dt = ±4 ≠ 0, so the tangent is vertical there (Topic 9.1) and d²y/dx² is not defined. The concavity changes at these points even though d²y/dx² is never zero. So when you look for changes in concavity on a parametric curve, check the values of t where dx/dt = 0 as well as where d²y/dx² = 0.

Common misconceptions

  • “d²y/dx² = (d²y/dt²) ÷ (d²x/dt²).” This shortcut is wrong. It fails for x = 2t, y = t², and gives the wrong sign in Worked example 1.
  • Stopping after step 3. d/dt (dy/dx) is not the answer. You must divide by dx/dt again.
  • Differentiating dy/dx with respect to x when it is written in t. You cannot treat t as x. Differentiate with respect to t, then convert by dividing by dx/dt.
  • Reading the concavity from d/dt (dy/dx). When dx/dt < 0 its sign is the opposite of the concavity, as in Worked example 2.
  • Not simplifying dy/dx first. A messy quotient makes step 3 long and error-prone. Cancel common factors and use identities such as sin 2t = 2 sin t cos t.
  • Ignoring where dx/dt = 0. Neither dy/dx nor d²y/dx² is defined there by this method, yet the concavity can change there, as in Worked example 3.
  • Mixing up over and under. Concave down means the tangent line is above the curve, so the estimate is too big.

Where this leads

Topic 9.3 uses dx/dt and dy/dt to find the arc length of a parametric curve. Topics 9.4 to 9.6 treat the parameter as time: there, the second derivatives x″(t) and y″(t) are the components of acceleration, which is a different quantity from d²y/dx². Keep the two separate. Next: Finding Arc Lengths of Curves Given by Parametric Equations (9.3).

Try the practice questions now, then use the revision notes and the checklist to consolidate.

Quick questions

Do Calculus AB students need this?

No. Parametric second derivatives are BC-only content. AB students can skip this page.

Why do I divide by dx/dt a second time?

dy/dx is a function of t, but d²y/dx² asks how it changes with x. The chain rule converts a t-derivative into an x-derivative by dividing by dx/dt, exactly as in Topic 9.1.

Do I need the formula (x′y″ − y′x″)/(x′)³?

No. It is an optional check that gives the same answer. The method of differentiating dy/dx and dividing by dx/dt always works and is easier to remember.

AP® and Advanced Placement® are trademarks registered by the College Board, which is not affiliated with, and does not endorse, this website. Marlbridge is an independent tuition provider. It is not an authorized AP course provider, an AP school or an exam centre, and these resources are not official College Board materials. Reading them or doing simulations does not meet the laboratory requirement of an AP science course.