In Integral Calculus, 201-SN3-RE (formerly 201-NYB), this chapter splits into three families: powers of trigonometric functions, handled by the parity of the exponents; trigonometric substitutions, called for by a square root; and inverse functions, which show up as antiderivatives. Each one has its recognition criterion, and it is that criterion, more than the technique, that you need to make automatic.
This sheet gives the three reference triangles, the decision table by parity, and the list of final steps that exam papers skip: the return to x, the change of bounds and the absolute value.
Mark this sheet as read or add it to your favourites: a free account, no password, keeps your read sheets and favourites from one visit to the next, tells you which chapter to tackle next and lets you ask for the method or exercise you are missing. Create your space, an email is enough.
The thread of the chapter
A trigonometric substitution is not finished when the integral in θ has been computed: you still have to come BACK to x, and that always goes through the right triangle. It is this last step that exam papers leave out.
•∫sinmxcosnxdx: if one exponent is ODD, split off one factor from it and substitute the OTHER function.
•If BOTH exponents are even, no substitution works: you must REDUCE THE POWERS with sin2x=21−cos2x and cos2x=21+cos2x.
•A useful shortcut: sinxcosx=2sin2x, which saves a whole substitution.
•Tangent-secant family: if the power of sec is EVEN, split off sec2x and set u=tanx; if the power of tan is ODD, split off secxtanx and set u=secx.
•Four antiderivatives to know: ∫tanxdx=−ln∣cosx∣, ∫cotxdx=ln∣sinx∣, ∫secxdx=ln∣secx+tanx∣, ∫cscxdx=−ln∣cscx+cotx∣.
sin2x and cos2x oscillate around 21 at twice the frequency: that is what 21∓cos2x says, and it is why even powers are reduced.
The cotangent-cosecant family is the exact mirror image of the tangent-secant family, up to sign: every name that starts with “co” carries a minus sign in its derivative, and therefore in its antiderivative.
The three substitutions and their triangles
•a2−x2 calls for x=asinθ, with dx=acosθdθ and a2−x2=acosθ.
•a2+x2 calls for x=atanθ, with dx=asec2θdθ and a2+x2=asecθ.
•x2−a2 calls for x=asecθ, with dx=asecθtanθdθ and x2−a2=a∣tanθ∣, the absolute value being mandatory.
•The reference TRIANGLE is what gets you back: on it you read sinθ, cosθ and tanθ in terms of x and a.
•A quadratic always comes down to one of the three forms by COMPLETING THE SQUARE: x2−6x+13=(x−3)2+4.
Three roots, three triangles: each time, the quantity under the root takes the side that is left once x and a are placed. This drawing is what gives the return to x at the end of the computation.
Draw the triangle once and for all at the start of the question: it gives at the same time the substitution, the simplification of the root and the return to x. Skipping it costs you all three.
Inverse functions, as antiderivatives and as integrands
•The absolute value in the third is not decoration: without it, the antiderivative is wrong on the negative branch.
•The square root of a DIFFERENCE (a constant minus a square) in the DENOMINATOR gives an inverse sine; a sum with no root gives an inverse tangent. That is the only criterion to remember.
•Integrating an inverse function is done by PARTS with the invisible factor 1: ∫arcsinxdx=xarcsinx+1−x2 and ∫arctanxdx=xarctanx−21ln(1+x2).
•The arccos and arccot versions differ only by a sign: the sum of the two antiderivatives of a pair is always 2πx.
arcsin exists only on [−1,1]; arctan is defined everywhere. Both stay between −2π and 2π.
Two antiderivatives that look different can both be right: they then differ by a CONSTANT. The check is to differentiate them, not to compare them term by term.
The mistakes that cost marks
These are the errors I correct most often in session. Each one costs marks on a paper, even when the reasoning behind it is right.
1.Not coming back to x after a trigonometric substitution
3 marks, and an answer in a variable that does not exist in the question
What not to write
“∫16−x2dx=∫dθ=θ+C.”
What to write
“You must come back to the original variable: x=4sinθ gives sinθ=4x, so θ=arcsin4x and the antiderivative is arcsin4x+C.”
Why: The variable θ is a computing tool that we introduced: it has no meaning in the question. An antiderivative must be expressed in the variable of the question.
2.Keeping the bounds in x after switching to θ
the whole question, 5 marks
What not to write
“∫0216−x2x2dx=∫0216sin2θdθ.”
What to write
“The bounds follow the variable: x=0 gives θ=0 and x=2 gives θ=6π, so ∫0π/616sin2θdθ.”
Why: Evaluating an antiderivative in θ at bounds expressed in x makes no sense. The alternative is to come back to x before evaluating, but you have to choose one of the two routes.
3.Writing x2−a2=atanθ without an absolute value
3 marks, and an antiderivative that is wrong on half the domain
What not to write
“With x=asecθ: x2−a2=atanθ.”
What to write
“x2−a2=a∣tanθ∣: the root is always positive, whereas tanθ changes sign depending on the branch.”
x2−a2 lives on TWO branches: on the left, x=asecθ gives tanθ<0, hence the absolute value.
Why: The function x=asecθ has two branches, x≥a and x≤−a, and tanθ has opposite signs on them. The case x≥a is the only one where the absolute value can be dropped, and you have to say so.
4.Attacking two even exponents with a substitution
the whole question, 5 marks, and ten minutes in a dead end
What not to write
“∫sin2xcos2xdx: I split off a cosx and set u=sinx.”
What to write
“Splitting off a cosine leaves cos1x, which you can do nothing with: you must REDUCE THE POWERS. With sinxcosx=2sin2x, the integral becomes ∫4sin22xdx=8x−32sin4x.”
Why: A substitution only works if the factor split off really is the derivative of the variable you set. With two even exponents, that factor does not exist, and no manipulation will make it appear.
5.Substituting on a quadratic without completing the square
the whole question, 5 marks
What not to write
“∫x2−6x+13dx: the form is x2−a2, so I set x=asecθ.”
What to write
“First complete the square: x2−6x+13=(x−3)2+4. So this is the form u2+a2 with u=x−3 and a=2, which calls for u=2tanθ.”
The quadratic never drops below 4: it never vanishes, so it does not factor. Its real form is a SQUARE PLUS a constant, which forces the tangent substitution.
Why: The quadratic has a negative discriminant: it does not factor, so it cannot be of the form x2−a2. Completing the square reveals its real form, and with it the right substitution.
6.Forgetting the dx of the trigonometric substitution
the whole question, 5 marks
What not to write
“x=4sinθ, so ∫16−x2x2dx=∫4cosθ16sin2θdθ.”
What to write
“dx=4cosθdθ: the integral becomes ∫4cosθ16sin2θ⋅4cosθdθ=∫16sin2θdθ.”
Why: It is precisely the dx that cancels the cosine in the denominator, and therefore the whole point of the substitution. Without it, the integral you get is harder than the one you started with.
7.Mixing up the criterion for the inverse sine and the one for the inverse tangent
3 marks
What not to write
“∫9+x2dx=arcsin3x.”
What to write
“With no root in the denominator, it is an inverse tangent: ∫9+x2dx=31arctan3x.”
With a root, 9−x21 exists only between −3 and 3: inverse sine. Without a root, 9+x21 is defined everywhere: inverse tangent.
Why: The only criterion is the presence of a root: a2−u21 gives an inverse sine, a2+u21 an inverse tangent. The factor a1, by the way, appears only in the second.
8.Believing that two antiderivatives that look different cannot both be right
no marks lost, but time and confidence
What not to write
“The answer key gives −4cos2x and I found 2sin2x: one of the two is wrong.”
What to write
“2sin2x=41−cos2x=−4cos2x+41: the two differ by a CONSTANT, so both are right.”
2sin2x and −4cos2x have the same shape, shifted by 41: two correct antiderivatives of sinxcosx.
Why: An antiderivative is only defined up to a constant. The only valid check is to differentiate both expressions: if the derivatives agree, both answers are correct.
9.Using a trigonometric substitution when a simple substitution is enough
no marks, but five to ten minutes lost on the exam
What not to write
“∫16−x2xdx: I set x=4sinθ.”
What to write
“The derivative of 16−x2 is −2x, which is already there: u=16−x2 gives −16−x2+C in three lines.”
Why: Trigonometric substitution is the last resort, not the first reflex. A factor x in front of the root almost always signals an ordinary substitution.
10.Getting the sign wrong in the cotangent-cosecant family
2 marks, but the error repeats over a whole question
What not to write
“∫csc2xdx=cotx.”
What to write
“∫csc2xdx=−cotx: everything that starts with “co” carries a minus sign in its derivative, and therefore in its antiderivative.”
Why: The “co” rule is systematic: (cos)′=−sin, (cot)′=−csc2, (csc)′=−csccot. It turns every formula of the tangent-secant family into its twin with a sign.
Which method to choose
Which technique for which trigonometric form
What you see under the integral sign
If an ODD exponent on sin or cos → split off one factor from the function with the odd exponent and substitute the other one
Example: ∫sin3xcos2xdx with u=cosx
If BOTH exponents even → REDUCE THE POWERS with the double-angle formulas; no substitution works
Example: ∫sin2xcos2xdx=8x−32sin4x
If an EVEN power of sec → split off sec2x and set u=tanx
Example: ∫tan3xsec4xdx
If an ODD power of tan → split off secxtanx and set u=secx
Example: ∫tan3xsec3xdx
If neither of the two cases above → convert with tan2x=sec2x−1 and start again
this is the road that leads to ∫sec3xdx, the cyclic case of the family
Parity is read on the exponent, not on the function: it is the only criterion, and it separates the three cases with no ambiguity at all.
Which substitution for which root
The form of the expression under the root, AFTER completing the square if it is a quadratic
If a2−u2 → u=asinθ, the root becomes acosθ
Example: 16−x2 with x=4sinθ
If a2+u2 → u=atanθ, the root becomes asecθ
Example: (x−3)2+4 with x−3=2tanθ
If u2−a2 → u=asecθ, the root becomes a∣tanθ∣
Example: x2−9 with x=3secθ
If a factor u appears in front of the root → ORDINARY substitution t= what is under the root, much faster
Example: ∫x16−x2dx
If the expression under the root is a quadratic → complete the square FIRST, then apply one of the three rows above
after completing the square, a negative squared term gives the sine form; with a positive squared term, the sign of the constant decides between the tangent form and the secant form
Before any trigonometric substitution, check whether the derivative of what is under the root is already there: in that case an ordinary substitution does the job in three lines instead of fifteen.
How the answer is expected to be written
A marker ticks steps. Here they are in order, with the concluding sentence expected word for word.
Carrying a trigonometric substitution through to the end
When to use it: The integrand contains a root of the form a2±u2 or u2−a2, and the derivative of what is under the root does not appear.
1Identify the form and set the matching substitution, writing x, dx AND the simplification of the root on three separate lines.
2DRAW the reference triangle right away: it will be used for the return, and it saves you from hunting for identities later.
3If the integral is definite, compute the new bounds in θ and say so; otherwise, plan the return to x.
4Rewrite the integral entirely in θ, check that no x is left, then compute it.
5Come back to x through the triangle, or evaluate directly at the bounds in θ if they were changed.
6Check by differentiating, or by testing a numerical value.
Concluding sentence
“Let x=4sinθ, so dx=4cosθdθ and 16−x2=4cosθ; the bounds x=0 and x=2 become θ=0 and θ=6π, and the integral reduces to ∫0π/616sin2θdθ.”
The trap: Skipping the triangle on the grounds that you will recall the identities from memory. That works for sinθ, not for tanθ and secθ, and that is where exam papers stop.
Marking: Typically 2 marks for the complete substitution, 1 for the bounds or the triangle, 3 for the computation in θ, 2 for the return or the evaluation, 1 for the conclusion.
Handling a power of sine and cosine
When to use it: The integrand is a product of powers of sin and cos.
1Note the two exponents and determine their parity, in writing.
2If one is odd, split off ONE factor of that function and convert the rest with sin2+cos2=1.
3Set the substitution on the OTHER function, and check that the factor split off really is its derivative, up to sign.
4If both are even, replace each square by its double-angle formula and expand, starting over once if needed.
5Integrate term by term, then come back to x if a substitution was made.
Concluding sentence
“The exponent of the sine is odd: split off sinx and write sin2x=1−cos2x; the substitution u=cosx, whose derivative −sinx is exactly the factor split off, gives a polynomial integral in u.”
The trap: Splitting off a factor from the function with the EVEN exponent. What is left then contains an odd power of the other function, which cannot be converted into squares, and the substitution fails.
Check before you hand in
Five minutes of checking recover more marks than one more problem started in a hurry.
No θ in the final answer
Reread the antiderivative: if it still contains θ, the return to x has not been made. It is the fastest check in the whole chapter.
An answer “θ+C”: you had to write arcsin4x+C.
Bounds consistent with the variable
Check that the bounds of the final integral are in the same variable as the integrand. Bounds in x on an integral in θ are a visible contradiction.
An integral ∫02sin2θdθ: the bound 2 is not an angle of the problem.
The derivative of the result
Differentiate the antiderivative you found and compare it with the integrand. This validates the substitution, the return and the constants all at once.
A derivative equal to the expected integrand divided by 4cosθ: the dx of the substitution was forgotten.
The absolute value where it is required
On x2−a2 and on the inverse secant, check that the absolute value is there and state which branch you are working on.
An antiderivative without an absolute value, correct for x>a and wrong for x<−a.
The order of magnitude of a definite integral
Compare the result with the area of a rough rectangle built on the interval and on an average value of the integrand.
A result of 0.72 for an integrand that rises from 0 to 1.15 on [0,2]: plausible.
The typical problem, taken apart
A definite integral by trigonometric substitution
Compute ∫0216−x2x2dx.
Also give the antiderivative expressed in x, so as to check the result by a second route.
The triangle carries the whole substitution: the hypotenuse is 4, the opposite side is x, and the adjacent side is exactly the root you want to simplify. It is what will give the return to x at the last step.
Step 1
The form 16−x2 is of type a2−x2 with a=4: set x=4sinθ, hence dx=4cosθdθ and 16−x2=4cosθ.
Why
The three lines are written together. No factor x stands in front of the root, so an ordinary substitution is ruled out and the trigonometric substitution is indeed the right one.
Step 2
New bounds: x=0 gives sinθ=0, so θ=0; x=2 gives sinθ=21, so θ=6π.
Why
Change the bounds right away, which will spare you the return to x for the numerical part. The substitution is increasing on [0,6π], so it is legitimate.
It is dx=4cosθdθ that cancels the cosine in the denominator: without it, a fraction would remain, and the whole point of the substitution would vanish.
Step 4
Both exponents are even: reduce the power. 16sin2θ=8(1−cos2θ), so the integral equals [8θ−4sin2θ]0π/6.
Why
Two techniques of the chapter are chained: the substitution has brought the problem down to an even power, which calls for power reduction. It is the most frequent case on final exams.
The lower bound makes both terms vanish, which makes the evaluation immediate. Keep the exact form before giving the decimal value.
Step 6
Antiderivative in x, through the triangle: sinθ=4x and cosθ=416−x2, so 8θ−8sinθcosθ=8arcsin4x−2x16−x2.
Why
First write sin2θ=2sinθcosθ so that the triangle is enough: without this identity, you would have to express a double angle, which cannot be read on the drawing.
Step 7
Check with the antiderivative: at x=2, 8arcsin21−2212=34π−23; at x=0, everything vanishes. The two routes agree.
Why
Two independent paths, one by changing the bounds and the other by coming back to x, give the same number: it is the best possible validation of a trigonometric substitution.
Step 8
Numerical check: the integrand is 0 at x=0 and 1.15 at x=2; an average rectangle of height 0.36 on a base of 2 gives 0.72.
Why
The order of magnitude confirms the result without any symbolic computation. A negative result, or one greater than 2.3, would have been rejected at once.
The conclusion, written out
“∫0216−x2x2dx=34π−23≈0.725, and the antiderivative is 8arcsin4x−2x16−x2+C.”
The classic mistake on this problem: Stopping at [8θ−4sin2θ] and evaluating between 0 and 2. The bounds are then values of x used as angles, and the result obtained, 16−4sin4≈19.0, has nothing to do with the area sought.
Learn by heart
•ODD exponent: split off one factor and substitute the other function. BOTH even: reduce the powers.
•sec to an even power: u=tanx. tan to an odd power: u=secx.
•A quadratic under a root: COMPLETE THE SQUARE before choosing the substitution.
•Root of a constant minus a square in the denominator: inverse sine. A sum with no root: inverse tangent, with a factor a1.
•Always finish with the RETURN to x, through the triangle, or by changing the bounds.
Frequently asked questions
How do you choose the trigonometric substitution?
Look at the form under the root. A difference of the type constant minus x squared calls for a sine. A sum calls for a tangent. A difference of the type x squared minus a constant calls for a secant. If the expression is a quadratic, first complete the square to reveal which of the three forms you really have.
What do you do when both exponents of sine and cosine are even?
Reduce the powers, that is, replace each square by its double-angle formula. No substitution can work in this case, because the factor you would split off would not be the derivative of the function you set. Power reduction turns the product into a sum of cosines of multiple angles, which integrates at once.
Why do you have to come back to x after a trigonometric substitution?
Because the angle you introduced is a computing tool that does not exist in the question. An antiderivative must be expressed in the variable of the question. The return is made with the reference right triangle, on which you read the sine, cosine and tangent of the angle in terms of x. For a definite integral, you can also change the bounds and avoid this return.
When is the absolute value necessary?
On the square root of x squared minus a squared, which becomes a times the absolute value of the tangent, and on the antiderivative involving the inverse secant. In both cases, the root is positive by definition while the trigonometric function changes sign depending on the branch. Stating the branch you are working on often lets you drop the absolute value.
How do you know whether an antiderivative gives an inverse sine or an inverse tangent?
Check whether there is a root in the denominator. The root of a difference gives an inverse sine, with no factor in front. A sum with no root gives an inverse tangent, with a factor equal to the reciprocal of the constant. That is the only criterion to remember, and it settles the question with no ambiguity.
Can two different antiderivatives both be right?
Yes, if they differ by a constant. An antiderivative is never unique. The only valid check is to differentiate both expressions and compare the derivatives with the integrand. Two answers that seem far apart can be turned into one another by a trigonometric identity.
A method is proved on a paper, not on a sheet. The set for the same chapter takes each of these traps into a problem, with the solution written out step by step.
Something missing from this sheet?A method, a trap, an exercise to practise it, 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.
Get in touch for a first session. We go through the chapter family by family, with a method for spotting in thirty seconds which technique applies, instead of trying them all.