In Lesson 12 you learned the chain rule: to differentiate a composite function, you multiply the derivative of the outside by the derivative of the inside. For example,
d/dx [(x² + 1)⁴] = 4(x² + 1)³ · 2x = 8x(x² + 1)³
Now run that backward. Suppose someone hands you the integral
∫ 8x(x² + 1)³ dx
You can see the "outside" piece (x² + 1)³ and a factor of 8x. Notice that 2x is exactly the derivative of the inside function x² + 1. That is not a coincidence — it is the fingerprint of a chain-rule derivative.
Your task: Without any new technique, just by reading the chain-rule example above in reverse, write down an antiderivative of 8x(x² + 1)³. Then differentiate your answer to check it. What inside function did you "undo," and how did its derivative show up in the integrand?
Every antiderivative rule you know is a derivative rule read backward. Integration by substitution (or u-substitution) is the reverse of the chain rule. When an integrand looks like a function of something, multiplied by the derivative of that something, substitution lets you collapse the whole expression into a single new variable and integrate easily.
The pattern to hunt for is:
∫ f(g(x)) · g'(x) dx
Here g(x) is an inside function and g'(x) is its derivative sitting in the integrand. If we let u = g(x), then du = g'(x) dx, and the integral becomes
∫ f(u) du
which has no composition left — just a plain antiderivative in u.
u. Pick the inside function — usually the quantity inside a power, an exponent, a denominator, or a trig argument — whose derivative also appears in the integrand (possibly off by a constant).du. Differentiate: du = u'(x) dx. Solve for dx or for the grouping you need.u. Every x and every dx must disappear. If a stray x survives, your choice of u was probably wrong.u.u = g(x) to return to x (indefinite integrals), and don't forget +C. For definite integrals you instead change the limits (more below).Find ∫ x(x² + 1)³ dx.
Choose u. The inside function is x² + 1. Let u = x² + 1.
Compute du. Then du = 2x dx. But the integral only has x dx, not 2x dx. No problem — solve for what we have:
du = 2x dx ⟹ x dx = (1/2) du
This is the constant-factor adjustment: a missing constant can always be supplied, because constants pass freely in and out of integrals. (A missing variable factor cannot — that would mean u was chosen incorrectly.)
Rewrite. Replace (x² + 1)³ with u³ and replace x dx with (1/2) du:
∫ x(x² + 1)³ dx = ∫ u³ · (1/2) du = (1/2) ∫ u³ du
Integrate.
= (1/2) · u⁴/4 + C = u⁴/8 + C
Back-substitute u = x² + 1:
∫ x(x² + 1)³ dx = (x² + 1)⁴ / 8 + C
Verify by differentiating (always do this):
d/dx [(x² + 1)⁴/8] = (1/8) · 4(x² + 1)³ · 2x = x(x² + 1)³ ✓
The derivative returns the original integrand exactly. That is your proof the antiderivative is correct.
∫ (g(x))ⁿ g'(x) dx = (g(x))^{n+1}/(n+1) + C (for n ≠ −1). Choose u = g(x).∫ e^{g(x)} g'(x) dx = e^{g(x)} + C. Choose u = g(x) (the exponent).∫ g'(x)/g(x) dx = ln|g(x)| + C. Choose u = g(x) (the denominator); the numerator is its derivative. The absolute value matters: ln|u| is the antiderivative of 1/u for all u ≠ 0.Trig composites work the same way: in ∫ sin²x cos x dx, the inside is sin x and cos x is its derivative, so u = sin x.
For a definite integral ∫ₐᵇ f(g(x)) g'(x) dx you have two valid options.
Method 1 — Change the limits of integration. When you substitute u = g(x), also convert the x-limits into u-limits. The lower limit x = a becomes u = g(a); the upper limit x = b becomes u = g(b). Then you never have to return to x.
Method 2 — Back-substitute to x first. Find the indefinite antiderivative in x (no +C needed for a definite integral), then evaluate at the original x-limits.
Both give the same number. Method 1 is cleaner and is what most AP scorers expect to see.
Evaluate ∫₀² x(x² + 1)³ dx.
Use u = x² + 1, so x dx = (1/2) du.
Method 1 — change limits. When x = 0, u = 0² + 1 = 1. When x = 2, u = 2² + 1 = 5. The integral becomes
∫₀² x(x² + 1)³ dx = (1/2) ∫₁⁵ u³ du = (1/2) · [u⁴/4]₁⁵
= (1/8)(5⁴ − 1⁴) = (1/8)(625 − 1) = 624/8 = 78
Method 2 — back-substitute. The antiderivative in x is (x² + 1)⁴/8, so
[(x² + 1)⁴/8]₀² = (5⁴)/8 − (1⁴)/8 = 625/8 − 1/8 = 624/8 = 78
Same answer, 78. Notice the #1 limit-change trap: if you change the limits to u-values, you must not also plug the original x-limits into a u-expression. Pick one method and stay consistent.
Problem. Find ∫ 3x²(x³ + 5)⁴ dx.
Strategy. The inside function is x³ + 5; its derivative is 3x², which appears exactly. Power pattern.
Solution. Let u = x³ + 5, so du = 3x² dx. The integral becomes
∫ u⁴ du = u⁵/5 + C = (x³ + 5)⁵/5 + C
Verify. d/dx [(x³ + 5)⁵/5] = (1/5)·5(x³ + 5)⁴·3x² = 3x²(x³ + 5)⁴ ✓
Because 3x² was already present, no constant adjustment was needed.
Problem. Find ∫ x e^{x²} dx.
Strategy. The exponent x² is the inside function; its derivative 2x is present up to the constant 2.
Solution. Let u = x², so du = 2x dx, giving x dx = (1/2) du.
∫ x e^{x²} dx = (1/2) ∫ e^u du = (1/2) e^u + C = (1/2) e^{x²} + C
Verify. d/dx [(1/2)e^{x²}] = (1/2)e^{x²}·2x = x e^{x²} ✓
g'/g → ln pattern (procedural) [NO CALC]Problem. Find ∫ 2x/(x² + 1) dx.
Strategy. The denominator x² + 1 has derivative 2x, which is exactly the numerator. This is the log pattern.
Solution. Let u = x² + 1, so du = 2x dx.
∫ 2x/(x² + 1) dx = ∫ (1/u) du = ln|u| + C = ln(x² + 1) + C
(We may drop the absolute value here because x² + 1 > 0 for all x, but write ln|u| as a habit.)
Verify. d/dx [ln(x² + 1)] = 1/(x² + 1) · 2x = 2x/(x² + 1) ✓
Problem. Evaluate ∫₀^{π/2} sin²x cos x dx.
Strategy. The inside function is sin x, and cos x (its derivative) is present. Change the limits.
Solution. Let u = sin x, so du = cos x dx. New limits: when x = 0, u = sin 0 = 0; when x = π/2, u = sin(π/2) = 1.
∫₀^{π/2} sin²x cos x dx = ∫₀¹ u² du = [u³/3]₀¹ = 1/3 − 0 = 1/3
Justification of the limit change. Because u = sin x is continuous and we substituted on a closed interval, evaluating the u-integral from u = sin 0 = 0 to u = sin(π/2) = 1 is equivalent to evaluating the original integral from x = 0 to x = π/2. The answer is 1/3.
Forgetting to convert dx via du. Students replace g(x) with u but leave a leftover dx, then "integrate" a mishmash. Why it's wrong: an integral in u must contain du, not dx. Fix: always write du = g'(x) dx explicitly and substitute it; if any dx or x survives, stop and recheck your u.
Dropping the constant factor. From du = 2x dx, students treat x dx as if it equals du. Why it's wrong: x dx = (1/2) du, not du — you lose a factor of 1/2. Fix: solve du for the exact grouping in the integrand and carry the constant out front.
On a definite integral, changing the limits AND back-substituting. Students convert the limits to u-values but then also plug the original x-limits into a u-expression (or vice versa). Why it's wrong: you end up evaluating at the wrong points. Fix: pick ONE method. If you change limits, evaluate the u-antiderivative at the u-limits and never return to x. If you back-substitute, use the original x-limits.
Forgetting +C on an indefinite integral. Why it's wrong: ∫ f dx is a family of antiderivatives; omitting +C loses the answer point on the AP exam. Fix: the instant you finish integrating an indefinite integral, write +C.
Choosing a u whose derivative isn't present. Students pick u = x² for ∫ (x² + 1)³ dx (note: no x factor). Why it's wrong: du = 2x dx, but there's no x to absorb, leaving a stray x. Fix: u must be an inside function whose derivative appears up to a constant — if it doesn't, substitution won't work directly.
Solutions in the Answer Key. Justification questions are marked ★.
∫ (2x + 1)⁵ dx =∫ x²(x³ − 1)⁴ dx =∫ e^{5x} dx =∫ x/(x² + 4) dx =∫ cos(3x) dx =∫ sec²x tan x dx =∫₀^{π/2} sin x cos x dx.∫₁^{e} (ln x)/x dx =∫₀² x e^{x²} dx = (round to three decimals)∫₀³ x/√(x² + 16) dx =Find ∫ (ln x)/x dx.
Evaluate ∫₀² x(x² + 1)³ dx using the change-of-limits method. Show the new limits.
★A student computes ∫₀² 2x(x²) dx by letting u = x² and writing ∫₀² u du = [u²/2]₀² = 2. Identify the error in the limits and give the correct value with justification.
★Explain why ∫ (x² + 1)³ dx cannot be evaluated by the substitution u = x² + 1 directly, but ∫ x(x² + 1)³ dx can. Reference what du requires.
Find ∫ x/√(x² + 16) dx (indefinite).
> Free Response Question (No calculator) — 9 points total
Water flows into a holding tank. For 0 ≤ t ≤ 4 hours, the rate at which water enters the tank is modeled by
R(t) = 6t/(t² + 1) liters per hour.
The tank contains 10 liters of water at time t = 0.
(a) (2 points) Find the rate at which water is entering the tank at time t = 4. Indicate units.
(b) (2 points) Find the time t in 0 < t < 4 at which R(t) attains its maximum value. Justify your answer.
(c) (3 points) Find the total amount of water, in liters, that enters the tank during the interval 0 ≤ t ≤ 4. Show the setup and use a substitution; give an exact value.
(d) (2 points) Find the total amount of water in the tank at time t = 4. Using the result of part (c), determine whether the average rate of flow over [0, 4] exceeds R(4), and interpret what that comparison means.
(a) Evaluate the model directly:
R(4) = 6(4)/(4² + 1) = 24/17 ≈ 1.412 liters per hour.
Water is entering at a rate of 24/17 ≈ 1.412 liters per hour at t = 4. (2 pts: 1 for value, 1 for units.)
(b) Differentiate using the quotient rule:
R'(t) = [6(t² + 1) − 6t(2t)] / (t² + 1)²
= [6t² + 6 − 12t²] / (t² + 1)²
= 6(1 − t²) / (t² + 1)².
R'(t) = 0 requires 1 − t² = 0, so t = 1 (rejecting t = −1, outside the domain). On (0, 1), 1 − t² > 0 so R' > 0; on (1, 4), 1 − t² < 0 so R' < 0. Since R' changes from positive to negative at t = 1, R has a relative maximum there, and because it is the only critical point in (0, 4), the maximum on the interval occurs at t = 1. (2 pts: 1 for t = 1, 1 for sign-change justification.)
(c) The total water entering over [0, 4] is the accumulation
∫₀⁴ R(t) dt = ∫₀⁴ 6t/(t² + 1) dt.
Let u = t² + 1, so du = 2t dt and 6t dt = 3 du. Change limits: t = 0 ⟹ u = 1; t = 4 ⟹ u = 17.
∫₀⁴ 6t/(t² + 1) dt = 3 ∫₁^{17} (1/u) du = 3[ln|u|]₁^{17}
= 3(ln 17 − ln 1) = 3 ln 17 liters.
So 3 ln 17 ≈ 8.500 liters enter the tank. (3 pts: 1 for the integral setup, 1 for a correct substitution/antiderivative, 1 for the exact value 3 ln 17.)
(d) The tank starts with 10 liters and gains 3 ln 17 liters, so at t = 4 it holds
10 + 3 ln 17 ≈ 18.500 liters.
The average rate of flow over [0, 4] is
(1/(4 − 0)) ∫₀⁴ R(t) dt = (3 ln 17)/4 ≈ 2.125 liters per hour.
Since 2.125 > 24/17 ≈ 1.412, the average rate of flow over the interval exceeds the instantaneous rate at t = 4. This means the flow was, on average, faster than its end-of-interval value — consistent with the rate peaking at t = 1 and then declining toward t = 4. (2 pts: 1 for 10 + 3 ln 17, 1 for the average-vs-R(4) comparison with interpretation.)
t = 1" with no sign analysis earns the value point but loses the justification point. The phrase that scores is "R' changes from positive to negative," not "it's a max."u = t² + 1, then evaluating 3[ln u] at 0 and 4 (the t-limits) instead of 1 and 17 — giving 3 ln 17 accidentally only if lucky, otherwise nonsense. Equally common: dropping the factor that makes 6t dt = 3 du.3 ln 17 as the tank total instead of the amount added. Read carefully: net change vs. total.1. (D) Let u = 2x + 1, du = 2 dx, so dx = (1/2)du. ∫ u⁵·(1/2)du = u⁶/12 + C = (2x + 1)⁶/12 + C. Check: d/dx[(2x+1)⁶/12] = 6(2x+1)⁵·2/12 = (2x+1)⁵ ✓. (A) forgets the 1/2; (C) forgets to divide by 6; (B) misapplies the power rule.
2. (B) u = x³ − 1, du = 3x² dx, so x² dx = (1/3)du. ∫ u⁴·(1/3)du = u⁵/15 + C = (x³ − 1)⁵/15 + C. Check: d/dx[(x³−1)⁵/15] = 5(x³−1)⁴·3x²/15 = x²(x³−1)⁴ ✓. (D) forgets 1/3; (A)/(C) mis-handle the constant.
3. (A) u = 5x, du = 5 dx, dx = (1/5)du. ∫ e^u·(1/5)du = (1/5)e^{5x} + C. Check: d/dx[(1/5)e^{5x}] = e^{5x} ✓. (B) forgets 1/5; (D) multiplies instead of dividing; (C) is nonsense.
4. (C) u = x² + 4, du = 2x dx, x dx = (1/2)du. ∫ (1/2)(1/u)du = (1/2)ln(x² + 4) + C. Check: d/dx[(1/2)ln(x²+4)] = (1/2)·2x/(x²+4) = x/(x²+4) ✓. (B) forgets 1/2; (A) doubles instead of halving; (D) treats it as a power, not a log.
5. (D) u = 3x, du = 3 dx, dx = (1/3)du. ∫ cos u·(1/3)du = (1/3)sin(3x) + C. Check: d/dx[(1/3)sin 3x] = cos 3x ✓. (B) forgets 1/3; (C) multiplies; (A) wrong sign.
6. (A) u = tan x, du = sec²x dx. ∫ u du = u²/2 + C = (1/2)tan²x + C. Check: d/dx[(1/2)tan²x] = tan x·sec²x ✓. (B) forgets 1/2; (D)/(C) choose the wrong u. (Equivalently (1/2)sec²x + C, differing only by the constant since sec²x = tan²x + 1.)
7. (C) u = sin x, du = cos x dx; limits 0 → 0, π/2 → 1. ∫₀¹ u du = [u²/2]₀¹ = 1/2. (B) is the integrand's value at an endpoint, not the integral; (A) forgets the 1/2; (D) is unrelated.
8. (B) u = ln x, du = (1/x)dx; limits x=1 → u=0, x=e → u=1. ∫₀¹ u du = [u²/2]₀¹ = 1/2. Antiderivative check: d/dx[(ln x)²/2] = (ln x)/x ✓. (A) forgets the 1/2; (D)/(C) ignore the substitution.
9. (C) u = x², du = 2x dx, x dx = (1/2)du; limits 0 → 0, 2 → 4. (1/2)∫₀⁴ e^u du = (1/2)(e⁴ − e⁰) = (e⁴ − 1)/2 ≈ 26.799. (D) 53.598 = e⁴ − 1 forgets the 1/2; (B) 1.000 confuses this with a sin·cos-type result; (A) miscomputes e⁴.
10. (D) u = x² + 16, du = 2x dx, x dx = (1/2)du; limits 0 → 16, 3 → 25. (1/2)∫₁₆²⁵ u^{−1/2} du = (1/2)[2√u]₁₆²⁵ = [√u]₁₆²⁵ = √25 − √16 = 5 − 4 = 1. Check antiderivative: d/dx[√(x²+16)] = x/√(x²+16) ✓. (C) reports √25 only; (A) reports √16 only; (B) is a stray 1/2.
11. u = ln x, du = (1/x)dx. ∫ u du = u²/2 + C = (ln x)²/2 + C. Check: d/dx[(ln x)²/2] = (ln x)·(1/x) = (ln x)/x ✓.
12. u = x² + 1, x dx = (1/2)du. New limits: x = 0 ⟹ u = 1; x = 2 ⟹ u = 5. (1/2)∫₁⁵ u³ du = (1/2)[u⁴/4]₁⁵ = (1/8)(625 − 1) = 78.
13. ★ The error: after u = x² the limits must change from x-values to u-values, but the student kept 0 and 2. Correct limits are x = 0 ⟹ u = 0 and x = 2 ⟹ u = 4. With du = 2x dx, the integral ∫₀² 2x(x²) dx = ∫₀⁴ u du = [u²/2]₀⁴ = 8. Justification: "Because the substitution u = x² changes the variable of integration, the limits must be expressed in terms of u: u = 0² = 0 and u = 2² = 4. Evaluating ∫₀⁴ u du gives 8." (Verify directly: ∫₀² 2x³ dx = [x⁴/2]₀² = 8 ✓.)
14. ★ The substitution u = x² + 1 produces du = 2x dx, so to rewrite the integral in u we need a factor of x (which 2x dx supplies up to the constant 2). In ∫ x(x² + 1)³ dx that x factor is present, so x dx = (1/2)du absorbs it and the integral becomes (1/2)∫ u³ du — fully in u. In ∫ (x² + 1)³ dx there is no x factor, so after substituting we would be left with ∫ u³ · dx with no way to convert the lone dx into du (a stray x remains). Substitution requires g'(x) to appear in the integrand up to a constant; here it does not.
15. u = x² + 16, x dx = (1/2)du. (1/2)∫ u^{−1/2} du = (1/2)·2u^{1/2} + C = √(x² + 16) + C. Check: d/dx[√(x²+16)] = (1/(2√(x²+16)))·2x = x/√(x²+16) ✓.
CalcIQ · Lesson 30 of 35 · Unit 6 — Integration and Accumulation of Change
This lesson is study material for the AP® Calculus AB exam. AP® is a registered trademark of the College Board, which was not involved in the production of, and does not endorse, this product.
Accuracy review: Every antiderivative in this lesson was verified by differentiation, and every definite integral's change of limits was checked against direct back-substitution (results agree). Reviewed for mathematical correctness by Isaac.