This is the corrected exercise set on elementary integration for Differential Calculus, 201-SN2-RE (formerly 201-NYA), the chapter that closes the semester. Antiderivatives and the substitution rule are part of the course description, and you meet them well before integral calculus: the course framework for objective 00UN asks you to compute elementary integrals, and the last week of Calculus I is devoted to them. Part A sets up the table of derivatives read backwards, the constant C and the theorem that justifies it, rewriting products and quotients, the transcendental lines, then the chain rule played backwards. Part B works on adjusting a constant factor, the line that a missing x makes you cross, sorting between the three possible decisions, and two problems where you work back from an acceleration or a velocity.
Two calculation exercises open the set, as in textbook drill lists. Exercise 0.1 lines up twelve antiderivatives to read directly from the reversed table of derivatives: polynomials, negative and fractional powers such as 3x2, the logarithm line ln∣x∣, the exponential, sine, cosine, sec2x and secxtanx, arctangent and arcsine, then a product to expand, a quotient to divide and cot2x through an identity. Exercise 0.2 lines up ten antiderivatives by substitution, from (4x−1)3 to sin4xcosx, 1+x4x and 1−9x21. Each line has a value to find at the given point, checked as you type it.
This set is written for CEGEP students in the Science program (Sciences de la nature) taking Differential Calculus, 201-SN2-RE.
The common thread fits in one sentence: at this point of the course you have no definite integral, no area under a curve and no fundamental theorem, so an antiderivative is nothing more than the table of derivatives read from right to left, and the ONLY check available to you is to differentiate your answer. The other half of the chapter is the constant C: differentiation erases it, integration cannot guess which one to put back, and the initial condition settles it. Two thirds of the marks lost in this chapter come down to it.
Self-checking setType your answers and check them question by question: the page tells you right or wrong before you open the solution. With an account, every answer right on the first try earns points, and your right answers stay filled in when you come back.
•DEFINITION: F is an antiderivative of f on an interval if F′=f on that interval. The whole table of antiderivatives is the table of derivatives read from right to left, and nothing else.
•THE ONLY CHECK: differentiate your answer and get the integrand back. It takes ten seconds, it relies on no theorem you have not yet seen, and it catches the great majority of errors.
•POWER: ∫xndx=n+1xn+1+C for n=−1, and the empty slot is filled by ∫xdx=ln∣x∣+C. Rewrite roots and quotients as powers of x first.
•THE CONSTANT: on an INTERVAL, two antiderivatives of the same function differ by a constant, a consequence of the Mean Value Theorem. Two answers that look different can therefore both be right.
•NO PRODUCT RULE AND NO QUOTIENT RULE: faced with a product or a quotient, you REWRITE, expand, divide, or recognize a substitution. You never invent a rule.
•SUBSTITUTION, the chain rule backwards: look in the integrand for a function u AND its derivative u′ as a factor, up to a CONSTANT. Set u, write du, leave no x behind, integrate, then go BACK to x.
•THE LINE: a constant comes out of the integral, a function of x never does. ∫x4−x2dx takes thirty seconds, ∫4−x2dx is not part of this course.
•INITIAL CONDITION: each integration adds a constant, so working back from a to x calls for TWO of them. Each constant is fixed by substituting the corresponding data, and you then check that the formula gives back what the problem said.
Calculations: antiderivatives read from the table
Exercise 0.1: Integrating with the table: twelve direct antiderivatives
For each function, find an antiderivative F by reading the table of derivatives from right to left, without substitution, then evaluate the antiderivative taken with C=0 at the given point.
One method for all twelve lines: first rewrite the integrand as a SUM of terms the table handles (powers of x, expanding a product, dividing a quotient term by term, a trigonometric identity), then integrate term by term, and finally check by differentiating. The calculator, in radian mode, is only used for the final value.
a) f(x)=5x4−6x2+4x−3, then F(2).
b) f(x)=x38+x23, then F(2).
c) f(x)=6x−3x2, then F(4).
d) f(x)=3x2−x5, then F(−2).
e) f(x)=4ex−7, then F(1).
f) f(x)=3cosx−2sinx, then F(3π).
g) f(x)=2sec2x−secxtanx, then F(4π).
h) f(x)=1+x28−1−x25, then F(21).
i) f(x)=x2(x−3)2, then F(2).
j) f(x)=(x+2)(x−1), then F(4).
k) f(x)=x22x3+x−6, then F(2).
l) f(x)=cot2x, then F(4π).
Show the solution
Answers
a)F(x)=x5−2x3+2x2−3x; F(2)=18
b)F(x)=−x24−x3; F(2)=−25
c)F(x)=4x23−3x32; F(4)≈24.440
d)F(x)=x3−5ln∣x∣; F(−2)≈−11.466
e)F(x)=4ex−7x; F(1)=4e−7≈3.873
f)F(x)=3sinx+2cosx; F(3π)=233+1≈3.598
g)F(x)=2tanx−secx; F(4π)=2−2≈0.586
h)F(x)=8arctanx−5arcsinx; F(21)≈1.091
i)F(x)=5x5−23x4+3x3; F(2)=532
j)F(x)=2x2+32x23−2x; F(4)=316≈5.333
k)F(x)=x2+ln∣x∣+x6; F(2)=7+ln2≈7.693
l)F(x)=−cotx−x; F(4π)=−1−4π≈−1.785
The move that decides everything happens BEFORE integrating: write the integrand as a sum of lines from the table. There is no product rule and no quotient rule for antiderivatives, so a product is expanded, a quotient is divided, a root becomes a power, and cot2 goes through an identity.
a)
Term by term, the exponent goes up by one and you divide by the new exponent: F(x)=x5−2x3+2x2−3x. So F(2)=32−16+8−6=18.
b)
f(x)=8x−3+3x−2, so F(x)=8⋅−2x−2+3⋅−1x−1=−x24−x3. So F(2)=−1−23=−25. Pitfall: the exponent −3 goes UP to −2, it does not go down to −4, and you divide by −2, hence the minus signs.
c)
f(x)=6x21−2x−31, so F(x)=6⋅23x23−2⋅32x32=4x23−3x32. So F(4)=4×8−3316≈32−7.560=24.440. Pitfall: dividing by a fractional exponent means multiplying by its reciprocal.
d)
The logarithm line handles the term in x1, the only power the power rule refuses: F(x)=x3−5ln∣x∣. So F(−2)=−8−5ln2≈−11.466. Pitfall: without the absolute value, ln(−2) does not exist.
e)
ex is its own antiderivative: F(x)=4ex−7x. So F(1)=4e−7≈3.873. Pitfall: treating ex as a power of x and writing x+1ex+1.
f)
∫cosxdx=sinx and ∫sinxdx=−cosx: F(x)=3sinx+2cosx. So F(3π)=233+1≈3.598. Pitfall: the sign, −2sinx has antiderivative +2cosx.
g)
sec2x is the derivative of tanx, and secxtanx that of secx: F(x)=2tanx−secx. So F(4π)=2−2≈0.586.
h)
The two inverse trigonometric lines: F(x)=8arctanx−5arcsinx, on (−1,1). So F(21)=8arctan21−65π≈3.709−2.618=1.091.
i)
Expand: x2(x−3)2=x2(x2−6x+9)=x4−6x3+9x2, so F(x)=5x5−23x4+3x3. So F(2)=532−24+24=532=6.4. Pitfall: integrating each factor separately and then multiplying, which makes no sense.
j)
Expand: (x+2)(x−1)=x−x+2x−2=x+x−2, so F(x)=2x2+32x23−2x. So F(4)=8+316−8=316≈5.333.
k)
Divide term by term: x22x3+x−6=2x+x1−x26, so F(x)=x2+ln∣x∣+x6. So F(2)=4+ln2+3≈7.693. Pitfall: ∫−6x−2dx=−6⋅−1x−1=+x6, the two minus signs cancel.
l)
No line handles cot2, but 1+cot2x=csc2x gives cot2x=csc2x−1, and csc2x is the negative of the derivative of cotx: F(x)=−cotx−x. So F(4π)=−1−4π≈−1.785. Pitfall: forgetting the minus sign in (cotx)′=−csc2x.
Check: DIFFERENTIATE each F you found and you must get exactly f back; on the calculator, the rate 0.002F(a+0.001)−F(a−0.001) must give f(a) to within a thousandth. A discrepancy by a constant factor betrays a forgotten division, a discrepancy in sign a trigonometric line read the wrong way.
Tick the exercises you have done or want to review: a free account, no password, keeps your ticks and your right answers from one visit to the next, tells you which chapter to tackle next and lets you ask for the exercise you are missing. Create your space, an email is enough.
Calculations: antiderivatives by substitution
Exercise 0.2: Integrating by substitution: ten direct antiderivatives
For each function, find an antiderivative F by substitution, then evaluate the antiderivative taken with C=0 at the given point.
One method for all ten lines: spot the inner function u whose derivative u′ is present as a factor, up to a constant; write du=u′(x)dx and adjust the constant; integrate in u with a line of the table; go back to x before evaluating.
a) f(x)=(4x−1)3, then F(1).
b) f(x)=x(x2−1)4, then F(2).
c) f(x)=x2x3+1, then F(2).
d) f(x)=x2+x+32x+1, then F(1).
e) f(x)=xe−x2, then F(1).
f) f(x)=sin(3x+1), then F(0).
g) f(x)=sin4xcosx, then F(2π).
h) f(x)=tan2(2x)sec2(2x), then F(8π).
i) f(x)=1+x4x, then F(1).
j) f(x)=1−9x21, then F(61).
Show the solution
Answers
a)F(x)=16(4x−1)4; F(1)=1681
b)F(x)=10(x2−1)5; F(2)=10243
c)F(x)=92(x3+1)23; F(2)=6
d)F(x)=ln(x2+x+3); F(1)=ln5≈1.609
e)F(x)=−21e−x2; F(1)=−2e1≈−0.184
f)F(x)=−31cos(3x+1); F(0)=−3cos1≈−0.180
g)F(x)=5sin5x; F(2π)=51
h)F(x)=6tan3(2x); F(8π)=61≈0.167
i)F(x)=21arctan(x2); F(1)=8π≈0.393
j)F(x)=31arcsin(3x); F(61)=18π≈0.175
The move that decides everything: find u AND its derivative in the integrand. A missing constant is adjusted by dividing; a missing x can never be adjusted. And an answer left in terms of u is not an answer: you go back to x before evaluating.
a)
u=4x−1, du=4dx: ∫u34du=16u4, so F(x)=16(4x−1)4 and F(1)=1681=5.0625. Pitfall: 4(4x−1)4, without the factor 41 from the inside.
b)
u=x2−1, du=2xdx, so xdx=2du: 21∫u4du=10u5, so F(x)=10(x2−1)5 and F(2)=10243=24.3.
c)
u=x3+1, du=3x2dx, so x2dx=3du: 31∫u21du=31⋅32u23, so F(x)=92(x3+1)23 and F(2)=92×27=6.
d)
The numerator IS the derivative of the denominator: u=x2+x+3, du=(2x+1)dx, ∫udu=ln∣u∣. Since x2+x+3>0 (discriminant 1−12<0), F(x)=ln(x2+x+3) and F(1)=ln5≈1.609.
e)
u=−x2, du=−2xdx, so xdx=−2du: −21∫eudu, so F(x)=−21e−x2 and F(1)=−2e1≈−0.184. Pitfall: losing the minus sign that comes from u′=−2x.
f)
u=3x+1, du=3dx: 31∫sinudu=−31cosu, so F(x)=−31cos(3x+1) and F(0)=−3cos1≈−0.180. Pitfall: the minus sign of ∫sinudu AND the 31.
g)
u=sinx, du=cosxdx: ∫u4du=5u5, so F(x)=5sin5x and F(2π)=51.
h)
u=tan(2x), du=2sec2(2x)dx: 21∫u2du=6u3, so F(x)=6tan3(2x) and, since tan4π=1, F(8π)=61≈0.167. Pitfall: taking u=sec(2x), whose derivative 2sec(2x)tan(2x) is not a factor.
i)
x4=(x2)2: u=x2, du=2xdx, so 21∫1+u2du=21arctanu. So F(x)=21arctan(x2) and F(1)=8π≈0.393. Pitfall: looking for a logarithm, when u′=2x is not the derivative of 1+x4.
j)
9x2=(3x)2: u=3x, du=3dx, so 31∫1−u2du=31arcsinu. So F(x)=31arcsin(3x) and F(61)=31⋅6π=18π≈0.175.
Check: differentiate each F with the chain rule and you must get f back exactly, the derivative of the inside cancelling the adjustment constant. If the derivative gives f multiplied by a number, the adjustment is missing; if it gives an extra x, the substitution was not allowed.
Part A: the basics (/50)
Exercise 1: The table of derivatives, read from right to left
You have spent the semester filling in a table: a function on the left, its derivative on the right. This chapter asks for nothing new, it asks you to read that same table FROM RIGHT TO LEFT. An antiderivative of f is a function F such that F′=f, so every line of the table of derivatives gives you an antiderivative formula for free.
The power rule (xn)′=nxn−1 reverses into ∫xndx=n+1xn+1+C, valid for every n=−1. Before integrating anything, rewrite the integrand as POWERS of x: a quotient x51 and a root 3x2 cannot be recognized as long as they keep their original form.
a) Compute ∫(6x2−8x3+5)dx, ∫x5dx and ∫3x2dx, first rewriting each integrand as a power of x.
b) Compute ∫(x4−x24)dx. The two terms look alike: do they come from the same line of the table?
c) Check BY DIFFERENTIATION the third antiderivative in a) and the answer to b). Which differentiation rule do you use each time?
d) A student answers ∫x5dx=lnx5+C. Refute this in ten seconds, without computing the true antiderivative again.
Show the solution
Answers
a)2x3−2x4+5x+C; −4x41+C; 53x5/3+C
b)4ln∣x∣+x4+C, two different lines of the table
c)Both derivatives give back the integrand: power rule, then logarithm
d)False: (lnx5)′=x5 and not x51
a)
First integral: linearity lets you work term by term, and the reversed power rule gives ∫6x2dx=6⋅3x3=2x3, ∫−8x3dx=−8⋅4x4=−2x4 and ∫5dx=5x. Answer: 2x3−2x4+5x+C. Second integral: rewrite x51=x−5, then ∫x−5dx=−4x−4=−4x41+C. The exponent goes from −5 up to −4: it INCREASES by 1, even though it stays negative, and this is where one test in three writes −6. Third integral: 3x2=x2/3, so ∫x2/3dx=5/3x5/3=53x5/3+C. Dividing by 35 means multiplying by 53: the fraction flips, it is not copied as is.
b)
The two terms certainly do NOT come from the same line. The first, x4=4x−1, carries exactly the exponent the power rule forbids, since it would produce a division by n+1=0. The logarithm line handles it: ∫xdx=ln∣x∣+C, so ∫x4dx=4ln∣x∣+C. The second, −x24=−4x−2, carries an ordinary exponent: ∫−4x−2dx=−4⋅−1x−1=4x−1=x4. Answer: 4ln∣x∣+x4+C. The trap is visual: two quotients that look alike, two formulas with nothing in common. The absolute value in the logarithm is not a frill, it makes the formula valid for x<0 as well.
c)
For a): (53x5/3)′=53⋅35x2/3=x2/3=3x2, and you get the integrand back: this is the power rule, the one from the first week. For b): (4ln∣x∣+x4)′=x4−x24, using (ln∣x∣)′=x1 and (4x−1)′=−4x−2 ✓. Remember this move: in differential calculus you have no other means of checking, neither a fundamental theorem nor a numerical value to compare with. Differentiating your answer takes ten seconds and catches the great majority of exam errors.
d)
Differentiate the proposed answer: (lnx5)′=x55x4=x5, by the chain rule. But the integrand was x51. The two do not even look alike, so the answer is wrong, and you did not need to know the right one to say so. The student applied the logarithm line on seeing a quotient, when that line handles ONLY the exponent −1. The confusion typically costs three marks out of ten on the exam, and it is always spotted by the same check.
Exercise 2: The constant C, what differentiation erases
Differentiating destroys information: all the functions F(x)+C have the same derivative. Integrating must therefore give that information back, and nobody can guess which one: hence the +C, which is not a decoration but the trace of a whole family of parallel curves.
The figure shows three members of the family y=2x3/2+C. They have the same slope at every x, hence the same derivative, and only one of them will pass through a prescribed point.
a) Check that F(x)=tan2x and G(x)=sec2x are both antiderivatives of f(x)=2sec2xtanx on (−2π,2π). What constant separates them?
b) State the theorem guaranteeing that two antiderivatives of the same function on an INTERVAL differ by a constant. Which theorem of the course does it follow from?
c) Find the antiderivative of f(x)=4x3−x22 that equals 7 at x=1.
d) Solve y′=3x with y(4)=10. Which of the three curves in the figure is a solution? Check your answer in two ways.
Show the solution
Answers
a)Both are antiderivatives of f; G−F=1
b)On an INTERVAL, F′=G′ implies F−G constant; a consequence of the Mean Value Theorem
c)F(x)=x4+x2+4
d)y=2x3/2−6; none of the three curves drawn
a)
Differentiate both candidates. F′(x)=2tanx⋅sec2x by the power rule and the chain rule, since (tanx)′=sec2x. G′(x)=2secx⋅(secxtanx)=2sec2xtanx, since (secx)′=secxtanx. Both derivatives equal f, so both functions are indeed antiderivatives of f ✓. Yet they do not look alike. The identity sec2x−tan2x=1 settles the matter: G−F=1, the constant that separates them is exactly 1. Numerical check at x=4π: F=1 and G=2 ✓. Two answers that look very different can therefore both be right, and this is the most common source of disputes over a test.
b)
Theorem: if F′=G′ on an INTERVAL I, then F−G is constant on I. It follows from the Mean Value Theorem, seen in the middle of the semester: set H=F−G, so H′=0 everywhere on I, hence for any two points a<b of I there exists c between them such that H(b)−H(a)=H′(c)(b−a)=0, so H(b)=H(a) for every pair of points, that is, H is constant. The word INTERVAL carries all the weight: on R∖{0}, which is a union of two intervals, ln∣x∣+1 on the left and ln∣x∣+5 on the right have the same derivative without differing by a single constant.
c)
Integrate term by term: ∫4x3dx=x4 and ∫−2x−2dx=−2⋅−1x−1=x2. The family is F(x)=x4+x2+C. The condition F(1)=7 gives 1+2+C=7, so C=4 and F(x)=x4+x2+4. Check by differentiation: F′(x)=4x3−x22 ✓, and check at the point: F(1)=1+2+4=7 ✓. BOTH checks count: the first checks the integration, the second checks the constant, and a wrong constant leaves the first one perfectly green. As an extra check, F(2)=16+1+4=21.
d)
y′=3x=3x1/2 gives y=3⋅3/2x3/2+C=2x3/2+C. The condition y(4)=10 gives 2⋅8+C=10, so C=−6 and y=2x3/2−6. None of the three curves in the figure is a solution: they have C=−12, C=−3 and C=6, and the point (4,10) lies between the first two, exactly as the figure shows. First check, by differentiation: (2x3/2)′=2⋅23x1/2=3x ✓. Second check, with the point: 2⋅43/2−6=16−6=10 ✓. Extra check: y(9)=2⋅27−6=48. An equation of the form y′=f(x) with a condition is the simplest of differential equations, and the whole chapter does nothing but solve it.
Exercise 3: Rewrite first, integrate afterwards
There is no product rule and no quotient rule for integration. When the integrand is a product or a quotient, the only move available is to REWRITE it as a sum of terms the table handles directly: expand, divide, or replace it with a trigonometric identity.
The figure shows the integrand of d), the parabola y=(2x−3)2. This integral can be computed in two perfectly legitimate ways, and the two answers do not look the same.
a) Compute ∫tan2xdx. No line of the table handles tan2: which identity solves the problem?
b) Compute ∫x2+1x2dx. The degree of the numerator equals that of the denominator: what must be done before anything else?
c) Compute ∫x(x+1)2dx for x>0.
d) Compute ∫(2x−3)2dx in two ways, by expanding and then by reversing the chain rule. Do the two answers contradict each other?
Show the solution
Answers
a)tanx−x+C, by tan2x=sec2x−1
b)x−arctanx+C, after dividing
c)x+4x+lnx+C
d)34x3−6x2+9x+C and 6(2x−3)3+C′: they differ by −4.5, both are right
a)
The table does not contain tan2, but it does contain sec2, which is the derivative of tan. The identity 1+tan2x=sec2x gives tan2x=sec2x−1, and the integral becomes a sum of two known lines: ∫(sec2x−1)dx=tanx−x+C. Check by differentiation: (tanx−x)′=sec2x−1=tan2x ✓. Numerical check at x=4π: the antiderivative equals 1−4π≈0.21. The trap in this question is wanting to integrate tan2 as a power, which would give 3tan3x: this answer is refuted in one differentiation, because the chain rule would bring in a stray sec2x.
b)
A quotient whose numerator has the same degree as the denominator is divided first. Here the division is immediate: x2+1x2=x2+1(x2+1)−1=1−x2+11. Both pieces are then lines of the table, the second being the arctangent line: ∫x2+1x2dx=x−arctanx+C. Check: (x−arctanx)′=1−x2+11=x2+1x2 ✓. At x=1 the antiderivative equals 1−4π≈0.21. What you must NOT write is a quotient of integrals: ∫vu=∫v∫u, and differentiating this false result shows it in one line.
c)
Expand the square before anything else: (x+1)2=x+2x+1, then divide each term by x: xx+2x+1=1+2x−1/2+x1. Each piece is now a line of the table: ∫1dx=x, ∫2x−1/2dx=2⋅1/2x1/2=4x and ∫xdx=lnx, the logarithm needing no absolute value since x>0. Answer: x+4x+lnx+C. Check: 1+2x4+x1=1+x2+x1 ✓, which is indeed the expansion obtained. At x=1 and with C=0, the antiderivative equals 1+4+0=5.
d)
First way, expand: (2x−3)2=4x2−12x+9, so the integral equals 34x3−6x2+9x+C. Second way, reverse the chain rule: the inside 2x−3 has derivative 2, so ∫(2x−3)2dx=3⋅2(2x−3)3+C′=6(2x−3)3+C′. The two answers do not contradict each other: expanding the second, 68x3−36x2+54x−27=34x3−6x2+9x−4.5. So they differ by the constant −4.5, which is allowed and expected, since two antiderivatives of the same function on an interval differ only by a constant. A marker who sees an unexpected answer differentiates it: that is the only arbiter.
Exercise 4: Transcendental functions, backwards
The second half of the table of derivatives is transcendental: exponential, logarithm, sine and cosine, tangent and secant, arcsine and arctangent. Reversed, it gives as many antiderivative formulas, and this is where SIGN errors pile up.
Remember the asymmetry: (sinx)′=cosx, so ∫cosxdx=sinx+C, but (cosx)′=−sinx, so ∫sinxdx=−cosx+C. A minus sign appears on one side and not on the other.
a) Compute ∫(5ex−1−x23)dx.
b) Compute ∫(4sinx−sec2x)dx, then give the value of this antiderivative at x=0 with C=0.
c) Compute ∫(1+x26−x2)dx, then check the answer by differentiation.
d) A student answers ∫4sinxdx=4cosx+C. Differentiate this answer and say which line of the table the student read the wrong way.
Show the solution
Answers
a)5ex−3arcsinx+C on (−1,1)
b)−4cosx−tanx+C; the antiderivative equals −4 at x=0
c)6arctanx−2ln∣x∣+C; it equals 23π≈4.71 at x=1
d)False: (4cosx)′=−4sinx, the right answer is −4cosx+C
a)
Two lines of the table, taken separately. For the first, (ex)′=ex is the only function that is its own derivative, so ∫5exdx=5ex. For the second, 1−x21 is exactly the derivative of arcsinx, so ∫−1−x23dx=−3arcsinx. Answer: 5ex−3arcsinx+C, on (−1,1) where the radical is defined. Check: 5ex−1−x23 ✓. The classic trap is to confuse 1−x21, which gives an arcsine, with 1+x21, which gives an arctangent: the radical and the minus sign under the root tell them apart at a glance.
b)
For the first term, ∫sinxdx=−cosx, so ∫4sinxdx=−4cosx. For the second, sec2x is the derivative of tanx, so ∫−sec2xdx=−tanx. Answer: −4cosx−tanx+C. Check by differentiation: 4sinx−sec2x ✓. At x=0 with C=0: −4⋅1−0=−4. This value is a useful check, because a sign error on the cosine would turn it into +4: a single evaluation at a well-chosen point is often enough to separate a right answer from an answer of the opposite sign.
c)
The first term is the arctangent line: ∫1+x26dx=6arctanx. The second is the logarithm line: ∫−x2dx=−2ln∣x∣. Answer: 6arctanx−2ln∣x∣+C. Check by differentiation: (6arctanx)′=1+x26 and (−2ln∣x∣)′=−x2, so the sum gives back the integrand ✓. Numerical check at x=1: 6⋅4π−2ln1=23π≈4.71. The trap here is to integrate 1+x26 as a logarithm, on the grounds that there is a quotient: but ln(1+x2) has derivative 1+x22x, and the stray 2x gives the error away immediately.
d)
Differentiate the proposed answer: (4cosx)′=−4sinx. You get the negative of the integrand, so the answer is wrong, and the right one is −4cosx+C. The student read the line (sinx)′=cosx and reversed it mechanically, believing that sine gives cosine in both directions. The table is asymmetric: differentiating cosine produces a minus sign, so integrating sine produces one too. This error usually costs half the marks of the question, while the check that catches it takes one line and five seconds. When in doubt about a sign, evaluate at x=0: (−cosx) equals −1 there and is increasing, which is consistent with an integrand sinx that is positive just after 0.
Exercise 5: The chain rule backwards: seeing u and its derivative
You have differentiated composite functions for six weeks: set the inside u, differentiate the outside, MULTIPLY by u′. Substitution is exactly this move played backwards. You look in the integrand for an inner function u AND its derivative u′ as a factor; if both are there, the integral comes down to a line of the table.
The protocol fits in three lines: set u=…, write du=u′(x)dx, replace everything, and not a single x may remain; integrate in u, then go BACK to x. An answer left in terms of u is not an answer.
a) Compute ∫8x(x2+3)3dx, writing out u, du and the return to x.
b) Compute ∫2+sinxcosxdx. Why is the absolute value unnecessary in the logarithm?
c) Compute ∫xexdx for x>0.
d) Compute ∫1+x2arctanxdx, then check by differentiation.
Show the solution
Answers
a)u=x2+3; (x2+3)4+C
b)ln(2+sinx)+C; no absolute value, the argument is at least 1
c)2ex+C
d)2(arctanx)2+C
a)
Set u=x2+3, so du=2xdx. The factor 8xdx is then 4du, and the integral becomes ∫4u3du=u4+C. Back to x: (x2+3)4+C. Check by differentiation: 4(x2+3)3⋅2x=8x(x2+3)3 ✓, you get exactly the integrand back. Numerical check at x=1 with C=0: 44=256. Note that you are not looking for u′ exactly, but for u′ UP TO A CONSTANT FACTOR: here 8x is 4 times 2x, and taking a constant out of the integral is always allowed.
b)
Set u=2+sinx, so du=cosxdx, and the numerator IS exactly du. The integral becomes ∫udu=ln∣u∣+C, that is ln(2+sinx)+C. The absolute value is unnecessary because sinx∈[−1,1], so 2+sinx∈[1,3]: the argument of the logarithm is always positive, and even bounded below by 1. Check: (ln(2+sinx))′=2+sinxcosx ✓. Numerical check at x=0: ln2≈0.69. Writing the absolute value would not be an error, only an unnecessary precaution that you must be able to justify.
c)
Set u=x, so du=2xdx, that is xdx=2du. The integral becomes ∫eu⋅2du=2eu+C, and going back to x gives 2ex+C. Check by differentiation: 2ex⋅2x1=xex ✓. Numerical check at x=1: 2e≈5.44. This is the most instructive case of the set: the derivative of the inside, 2x1, is hidden in a DENOMINATOR, and you only recognize it by remembering the derivative of the square root. An integrand that contains a function and a strange denominator is almost always a disguised substitution.
d)
Set u=arctanx, so du=1+x2dx, and once again the rest of the integrand IS exactly du. The integral becomes ∫udu=2u2+C, that is 2(arctanx)2+C. Check by differentiation: 22arctanx⋅1+x21=1+x2arctanx ✓. Numerical check at x=1: 2(π/4)2≈0.31. The reflex to build is this one: when 1+x21 is a factor, think of the arctangent not only as a result but as a CANDIDATE u. The same remark holds for x1, which calls for u=lnx, and for cosx, which calls for u=sinx.
Part B: problems and reasoning (/50)
Exercise 6: When a constant factor is missing, and when an x is missing
A constant comes out of the integral, a function of x never does. The whole difficulty of substitution lies on this line: if u′ is present up to a CONSTANT factor, you adjust and the integral falls into place; if an x is missing, the integral changes nature completely and often leaves the course.
The figure shows y=x4−x2, the integrand of a). Its neighbour y=4−x2, without the factor x, is a semicircle, and it is not handled the same way at all.
a) Compute ∫x4−x2dx, stating u and du, then give its value at x=0 with C=0.
b) Compute ∫sec2(7x)dx. Where does the coefficient in front of the tangent come from?
c) A student writes ∫4−x2dx=−31(4−x2)3/2+C. Differentiate this answer: what exactly do you get, and what is missing from the starting integrand?
d) Compute ∫x2+9xdx and check by differentiation.
Show the solution
Answers
a)−31(4−x2)3/2+C; it equals −38≈−2.67 at x=0
b)71tan(7x)+C, the 71 offsets the chain rule
c)Its derivative is x4−x2: an extra factor x comes out, and the integral in c) is not part of 201-SN2-RE
d)x2+9+C; it equals 5 at x=4
a)
Set u=4−x2, so du=−2xdx, that is xdx=−21du. The integral becomes −21∫u1/2du=−21⋅32u3/2=−31u3/2, and going back to x gives −31(4−x2)3/2+C. Check by differentiation: −31⋅23(4−x2)1/2⋅(−2x)=x4−x2 ✓. Value at x=0 with C=0: −31⋅43/2=−38≈−2.67. The factor −21 is a CONSTANT, so it passes through the integral without difficulty: this is the comfortable case, where u′ is present up to the coefficient.
b)
Set u=7x, so du=7dx and dx=7du. The integral becomes 71∫sec2udu=71tanu+C, that is 71tan(7x)+C. The coefficient 71 comes from the derivative of the inside: when you differentiate tan(7x), the chain rule brings out a 7, and the 71 is there to cancel it. Check: 71⋅7sec2(7x)=sec2(7x) ✓. Numerical check: at x=28π, tan4π=1 and the antiderivative equals 71≈0.1429. The most widespread error is to forget this coefficient and answer tan(7x)+C, which gives an integrand seven times too large.
c)
Differentiate: (−31(4−x2)3/2)′=−31⋅23(4−x2)1/2⋅(−2x)=x4−x2. So you get the integrand of a) back, not that of c): an EXTRA factor x comes out. This is exactly the error of believing you can manufacture the missing x. But ∫4−x2dx is a different matter: substitution does not apply, for lack of u′, and its antiderivative requires a trigonometric substitution, a technique of the integral calculus course that follows. Remember the line: without u′ present up to a constant, you do not substitute, you stop.
d)
Set u=x2+9, so du=2xdx and xdx=21du. The integral becomes 21∫u−1/2du=21⋅2u1/2=u, that is x2+9+C. Check by differentiation: (x2+9)′=2x2+92x=x2+9x ✓. Numerical check at x=4 with C=0: 25=5. Compare with c): here the x in the numerator is precisely what makes the substitution possible, whereas ∫x2+9dx, without this x, again leaves the course. A single factor x separates a thirty-second integral from an integral of another semester.
Exercise 7: Same numerator, four decisions
Faced with an integral, the question is not which technique to apply, it is which of the three decisions of the chapter to make: read the table directly, REWRITE the integrand, or SUBSTITUTE. The four integrals that follow are built on the same numerator and the same quadratic, and each one calls for a different decision.
The sorting criterion is always the same: is the numerator, up to a constant, the derivative of the denominator? If so, substitute; if not, see whether a rewriting leads back to the table.
a) Compute ∫x2x+1dx.
b) Compute ∫x2+2x+5x+1dx. Why is the absolute value superfluous here?
c) Compute ∫(x+1)ex2+2xdx.
d) Compute ∫x2+2x+5dx, first writing the denominator in vertex form (by completing the square). Which of the four integrals needed NO substitution?
Show the solution
Answers
a)ln∣x∣−x1+C
b)21ln(x2+2x+5)+C; the quadratic is at least 4, so no absolute value
c)21ex2+2x+C
d)21arctan2x+1+C; only the integral in a) needed no substitution
a)
The denominator is x2, whose derivative is 2x: the numerator x+1 is not a constant multiple of it, so substitution is ruled out. REWRITE: x2x+1=x1+x21=x−1+x−2. Two lines of the table follow: ∫x−1dx=ln∣x∣ and ∫x−2dx=−x1. Answer: ln∣x∣−x1+C. Check by differentiation: x1+x21 ✓, which is indeed the starting rewriting. Value at x=1 with C=0: 0−1=−1. This is the only one of the four integrals settled without setting any u.
b)
The denominator x2+2x+5 has derivative 2x+2=2(x+1): the numerator is therefore u′ up to a factor 21, and substitution is called for. Set u=x2+2x+5, du=(2x+2)dx, so (x+1)dx=21du. The integral becomes 21∫udu=21ln∣u∣+C, that is 21ln(x2+2x+5)+C. The absolute value is superfluous because x2+2x+5=(x+1)2+4≥4>0 for every x: the quadratic has discriminant 4−20=−16<0, it never vanishes. Check: 21⋅x2+2x+52x+2=x2+2x+5x+1 ✓. At x=0: 21ln5≈0.80.
c)
Here the exponent is x2+2x, whose derivative is 2x+2=2(x+1), and the factor in front of the exponential is precisely x+1: same structure as in b), same substitution. Set u=x2+2x, du=(2x+2)dx, so (x+1)dx=21du, and the integral becomes 21∫eudu=21eu+C, that is 21ex2+2x+C. Check by differentiation: 21ex2+2x⋅(2x+2)=(x+1)ex2+2x ✓. Value at x=0 with C=0: 21e0=0.5. The methodological point is that u is not chosen from the look of the integrand, but from whose derivative is available.
d)
The numerator is 1: it is not the derivative of the denominator, so the previous substitution does not work. Complete the square, x2+2x+5=(x+1)2+4, then aim for the arctangent line, ∫u2+1du. Set u=2x+1, so x+1=2u and dx=2du: the integral becomes ∫4u2+42du=21∫u2+1du=21arctanu+C, that is 21arctan2x+1+C. Check: 21⋅1+(2x+1)21⋅21=4+(x+1)21=x2+2x+51 ✓. At x=1: 21arctan1=8π≈0.39. The one that needed no substitution is the integral in a).
Exercise 8: Five statements to correct
Each of the five statements below was taken from a student's test. For each one, say whether it is true or false, correct it, and justify the correction WITH A DIFFERENTIATION or with a numerical counterexample.
1) “∫f(x)g(x)dx=(∫f(x)dx)(∫g(x)dx), just like the derivative of a product.”
2) “The constant C is a detail of notation: F(x)=x2 and F(x)=x2+7 are the same antiderivative.”
3) “∫xdx=lnx+C on all of R∖{0}.”
4) “∫(3x+2)5dx=6(3x+2)6+C, by the power rule.”
5) “Differentiating then integrating, or integrating then differentiating, comes to the same thing: you get f back in both cases.”
Show the solution
Answers
1)False: 3x3 equals 9 at x=3, 4x4 equals 20.25
2)False: two different antiderivatives of the same function
3)False: ln∣x∣+C, and one constant PER INTERVAL
4)False: 18(3x+2)6+C
5)False in one direction: ∫f′(x)dx=f(x)+C, the constant is lost
1)
FALSE, and doubly so. First, there is no product rule for integration; second, the product rule for differentiation does not have this form either, since (fg)′=f′g+fg′. Numerical counterexample with f(x)=g(x)=x: the left-hand side is ∫x2dx=3x3, which equals 9 at x=3 with C=0; the right-hand side is (2x2)2=4x4, which equals 20.25 at x=3. The two numbers differ, so the statement falls. Correction: faced with a product, you must REWRITE, or recognize a substitution, or wait for the integral calculus course and integration by parts.
2)
FALSE. These two functions have the same DERIVATIVE, 2x, but they are not the same function: their graphs are two distinct curves, shifted vertically by 7, and F(0) equals 0 in one case and 7 in the other. Correction: they are two different antiderivatives of the same function. The expression x2+C does not denote a function but a FAMILY of functions, and an extra condition, a prescribed point, selects exactly one of them. Forgetting the +C in a general answer costs one mark; getting C wrong in an initial value problem costs the whole problem.
3)
FALSE because of lnx, which is defined only for x>0, and false in its scope as well. Correction in two steps. First the absolute value: ∫xdx=ln∣x∣+C, and differentiating confirms that for x<0, (ln(−x))′=−x−1=x1 ✓. Then the scope: R∖{0} is not an INTERVAL but the union of two intervals, and the theorem on antiderivatives applies only on an interval. The general antiderivative there therefore calls for TWO constants, one per interval. In practice, you work on the interval where the problem lives and a single constant is enough.
4)
FALSE, a factor 31 is missing. The check shows it in one line: (6(3x+2)6)′=66(3x+2)5⋅3=3(3x+2)5, that is three times the integrand. Correction: ∫(3x+2)5dx=18(3x+2)6+C, the correct denominator being 18=6×3. This is the most frequent error of the chapter, and it comes from applying the power rule to a composite function without offsetting the derivative of the inside. The substitution u=3x+2, du=3dx, makes it impossible: the 31 appears by itself as soon as you write dx=3du.
5)
FALSE in one direction out of two. Integrating then differentiating gives back exactly f: dxd(∫f(x)dx)=f(x), the constant disappears when you differentiate. But differentiating then integrating does not give back f: ∫f′(x)dx=f(x)+C, and nothing says this C is zero. Example: start from f(x)=x2+7, differentiate to get 2x, integrate to get x2+C. The information 7 is lost, and no computation recovers it. Correction: differentiation ERASES the constant, and integration cannot guess which one to put back. That is precisely why a serious antiderivative problem always supplies an initial condition.
Exercise 9: Problem: two constants, and a condition that is not at t = 0
A cart leaves a platform and accelerates along a horizontal rail. A sensor gives its acceleration, a(t)=t+16 in metres per second squared, for t≥0 in seconds. The radar on the platform measured v(3)=30 metres per second, and the origin of positions is taken at the platform, so x(0)=0.
Working back from a to x takes TWO integrations, hence two constants, hence two conditions. Here the first condition is not given at time zero, which means it cannot be read directly off the formula.
a) Find v(t). Which piece of data fixes the first constant, and why can it not be applied before integrating?
b) Find x(t), then the position of the cart at t=3 seconds.
c) Check your x(t) by two successive differentiations, and check both conditions.
d) A student sets both constants equal to zero “since the cart starts from the platform”. What does he find at t=3, and by how many metres is he off?
e) If you knew only a(t) and x(0), could you find x(t)? Explain by counting the constants.
Show the solution
Answers
a)v(t)=12t+1+6, with C1=6 fixed by v(3)=30
b)x(t)=8(t+1)3/2+6t−8; x(3)=74 m
c)x′=v and v′=a ✓, with x(0)=0 and v(3)=30 ✓
d)He finds 64 m, that is 10 m short, and his v(3)=24 contradicts the radar
e)No: two integrations, two constants, so two conditions are needed
a)
Write a(t)=6(t+1)−1/2 and integrate with the reversed power rule: ∫6(t+1)−1/2dt=6⋅1/2(t+1)1/2=12t+1, the inside t+1 having derivative 1, which adds no coefficient. So v(t)=12t+1+C1. The data v(3)=30 fixes C1: 124+C1=24+C1=30, so C1=6 and v(t)=12t+1+6. The condition cannot be applied before integrating, because C1 only exists after the integration: the condition says nothing about a, it gives a value of v, and v is known only once the antiderivative is written. Check: v(0)=12+6=18 metres per second.
b)
Integrate v: ∫12(t+1)1/2dt=12⋅3/2(t+1)3/2=8(t+1)3/2, and ∫6dt=6t. So x(t)=8(t+1)3/2+6t+C2. The condition x(0)=0 gives 8⋅1+0+C2=0, so C2=−8 and x(t)=8(t+1)3/2+6t−8. Position at t=3: 8⋅43/2+18−8=8⋅8+10=74 metres. Notice that C2 is not zero even though the initial position is: the term 8(t+1)3/2 already equals 8 at t=0, and that is what the constant must offset. Confusing “zero initial position” with “zero constant” is the error in d).
c)
First differentiation: x′(t)=8⋅23(t+1)1/2+6=12t+1+6=v(t) ✓. Second differentiation: v′(t)=12⋅2t+11=t+16=a(t) ✓. You get the starting acceleration back, so both integrations are right. That leaves the conditions, which are a SEPARATE check: x(0)=8+0−8=0 ✓ and v(3)=12⋅2+6=30 ✓. The two checks are independent, and that is what makes them useful: an antiderivative error makes the first one fail, a constant error makes only the second one fail.
d)
With C1=C2=0, he gets v(t)=12t+1 and x(t)=8(t+1)3/2, so x(3)=8⋅8=64 metres instead of 74: he is off by 10 metres. His error can be detected without even knowing the right answer, since his v(3)=12⋅2=24 contradicts the radar measurement, which said 30. This is the check to make every time: once the constants are set, substitute EACH piece of data from the problem into the formula obtained. A wrong constant leaves the check by differentiation perfectly green, which is why it so often goes unnoticed.
e)
No. Each integration introduces a constant, so working back from a to x introduces two, and exactly TWO conditions are needed to fix them. With the single piece of data x(0)=0, you get x(t)=8(t+1)3/2+C1t+C2 with C2=−8−0, that is x(t)=8(t+1)3/2+C1t−8: a family of trajectories indexed by C1, that is, by the unknown initial velocity. Physically, this is obvious: an accelerometer cannot tell whether the cart was already moving. The same remark applies to a phone, which measures its acceleration without ever being able to deduce on its own where it is or how fast it is going.
Exercise 10: Problem: the drone, a substitution and a forgotten C
A drone takes off from the ground and climbs vertically. Its rate of climb is v(t)=t2+420t in metres per second, for t≥0 in seconds, and its altitude at takeoff is zero, z(0)=0.
The figure shows the shape of this velocity. It starts at zero, goes through a maximum, then decreases slowly: the drone climbs more and more slowly, which does not mean that it stops climbing.
a) Find z(t) by a substitution that you will state, then fix the constant.
b) At what altitude is the drone at t=6 seconds?
c) At what instant is the rate of climb at its maximum, and what is it then?
d) Check z(t) by differentiation. Does the drone eventually reach a limiting altitude?
e) A student forgets the constant and announces z(6)=10ln40. What altitude does he announce, by how much is he off, and what does he unwittingly claim about the moment of takeoff?
Show the solution
Answers
a)u=t2+4; z(t)=10ln4t2+4, with C=−10ln4≈−13.86
b)z(6)=10ln10≈23.03 m
c)t=2 s, with v(2)=5 m/s
d)z′=v ✓; no limiting altitude, z(t)→+∞
e)He announces 36.89 m, that is 13.86 m too much, and he places the takeoff at an altitude of 13.86 m
a)
The denominator t2+4 has derivative 2t, and the numerator 20t is its constant multiple 10×2t: substitution is called for. Set u=t2+4, so du=2tdt and 20tdt=10du. The integral becomes 10∫udu=10ln∣u∣+C, that is 10ln(t2+4)+C, the absolute value being unnecessary since t2+4≥4>0. The condition z(0)=0 gives 10ln4+C=0, so C=−10ln4 and z(t)=10ln(t2+4)−10ln4=10ln4t2+4. Writing the answer in this condensed form is a real gain: the initial condition can be read in it, since the logarithm equals ln1=0 at t=0.
b)
z(6)=10ln436+4=10ln10≈23.03 metres. You can check the order of magnitude without a calculator: the velocity exceeds 4 metres per second roughly between the first and the fourth second, then drops back towards 2, so about twenty metres in six seconds is plausible. This plausibility check is worth as much as the computation: an answer of 200 metres or 2 metres should have raised a flag. Notice also the pleasant coincidence, 440=10, which produces a natural logarithm of ten.
c)
The velocity is at its maximum where v′(t)=0. Differentiate with the quotient rule: v′(t)=20⋅(t2+4)2(t2+4)−t⋅2t=(t2+4)220(4−t2). The denominator is always positive, so the sign is that of 4−t2: positive for t<2, negative for t>2. The velocity increases then decreases, so t=2 seconds is a MAXIMUM, and it is absolute on [0,+∞) since it is the only critical number. Value: v(2)=840=5 metres per second. This question is pure differential calculus, asked in the middle of an integration problem: the two halves of the course answer each other.
d)
Check: z′(t)=10⋅t2+42t=t2+420t=v(t) ✓, differentiating ln(t2+4) with the chain rule, the −10ln4 being a constant with zero derivative. There is NO limiting altitude: t2+4→+∞, so z(t)→+∞. The drone climbs forever, more and more slowly, but without ever levelling off. A velocity that tends to zero does not imply a bounded altitude, and this is a common misreading. Order of magnitude: it would take about 109 seconds to pass 80 metres, the logarithm being formidably slow.
e)
He announces 10ln40≈36.89 metres, instead of 23.03: he is off by 10ln4≈13.86 metres, too high. And he unwittingly claims that the drone was already at an altitude of 10ln4≈13.86 metres AT THE MOMENT OF TAKEOFF, since his formula gives z(0)=10ln4. This is how a forgotten C is spotted: you evaluate your own answer at the instant where the problem gives the condition, and you see whether it says what the problem says. The check takes five seconds, and it saves the five marks this kind of problem puts on the constant.
The rate of climb is at its maximum at t=2 s, the only critical number of v on [0,+∞), where it equals v(2)=5 m/s.
Something missing from this set?A type of exercise your teacher gives, a problem you are stuck on, a mistake you spotted: tell me, I read every request.
Sending goes through your free space: that is how I can answer you and tell you when the exercise is added. An email is enough, no password. Create your space.
Rather not create an account? Write to me directly.
Looking for a differential calculus tutor in Montreal?
Get in touch for a first session. We work on elementary integration at the level of real end-of-semester tests, from the reversed table all the way to problems with two constants.