Differential Calculus 201-SN2-RE • CEGEP, Montreal
Revision sheet: derivatives of transcendental functions (201-SN2-RE)
This sheet does not rewrite the table of derivatives: you already have it in your notes, and copying a table neither loses nor earns marks. It deals with what really costs them, namely the factor u′ that disappears as soon as there are two layers, the power rule applied to 2x, cos(x2) confused with cos2x, the absolute value forgotten in arccos(x3), and the domain of the logarithm discovered after differentiating. It also connects with the tg, cotg and cosec notation of French-language Quebec textbooks, and names the trigonometric logarithms that simplify, such as ln(secx+tanx), whose derivative is secx.
It is written for CEGEP students in the Science program (Sciences de la nature) taking Differential Calculus, 201-SN2-RE (formerly 201-NYA). Once you have read the sheet, the corrected exercise set of the same chapter puts every reflex to the test.
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 transcendental derivative is never read straight off the table: there is always a factor u′ behind it, and a domain to check before differentiating. Almost every mark lost in this chapter comes from one or the other.
•Reading rule: the three functions whose name starts with “co” carry the MINUS sign. Nothing else needs memorizing about the signs.
•All of this assumes the argument is in RADIANS. In degrees, (sinx)′=180πcosx, which is never what the problem wants.
•Notation of French-language Quebec textbooks: tan is also written tg, cot cotg and csc cosec. Same functions, same derivatives: (cotgx)′=−cosec2x.
The slope of y=sinx (green tangent lines) is read on y=cosx (dashed): 1 at 0, 0 at 2π where the sine levels off. In radians only.
The three “co” derivatives can also be recovered by differentiating the cofunction: cosx=sin(2π−x) produces the minus sign through the chain rule, since the inner derivative is −1.
Exponential and logarithm: the base changes a FACTOR, never the form
•(ex)′=ex, and more generally (eu)′=u′eu.
•(ax)′=axlna: the factor lna is the only trace of the base.
•(lnx)′=x1 for x>0, (lnu)′=uu′, and (ln∣x∣)′=x1 for every x=0.
•(logax)′=xlna1, which is just (lnalnx)′ with the constant lna1 in front.
•Three logarithms that simplify, because u′ is a multiple of u: (lnsinx)′=cotx, (lnsecx)′=tanx and (ln(secx+tanx))′=secx.
The derivative of 2x, dashed, is the same curve multiplied by ln2, so it runs BELOW it: only the base e gives a curve that coincides with its derivative.
The number e is defined exactly by this: it is the only base whose tangent line at 0 has slope 1, hence the only base for which the factor lna equals 1 and disappears.
The factor u′: one per layer, no exceptions
•dxd[F(u(x))]=F′(u)⋅u′. The table gives F′; you are the one who writes u′.
•Three layers give three factors: (sin(e3x))′=cos(e3x)⋅e3x⋅3.
•Count check: as many multiplied factors as nested functions. Two nested functions and only two factors, one is missing.
•The only case where u′=1 is u(x)=x: that is the line of the table, and it is rare on a test.
•A radical is rewritten as a power BEFORE differentiating: 3tan2u=(tanu)32. The exponent −31 that comes out of the derivative says where it does not exist: where tanu=0.
Inverse and hyperbolic functions: the two tables where the sign is at stake
•(arcsinx)′=1−x21 and (arccosx)′=1−x2−1, both for ∣x∣<1 STRICTLY.
•(arctanx)′=1+x21 on all of R, with no condition.
•(sinhx)′=coshx and (coshx)′=sinhx, WITHOUT a minus sign, with cosh2x−sinh2x=1.
•Consequence to remember: cos and sin satisfy y′′=−y and oscillate; cosh and sinh satisfy y′′=y and shoot off to infinity.
cosh and sinh differentiate into each other, without a minus sign: y′′=y, and the curves shoot off to infinity. The cosine, with its minus sign, satisfies y′′=−y and oscillates.
The domain of the derivative is the INTERIOR of the domain of the function. arcsin is defined at 1, its derivative is not: the tangent line there is vertical.
The rules in table form
Each row reads from left to right: the assumptions, then the result. A red cell is not an answer, it is the finding that the form settles nothing and the instruction to rewrite it. Every case is followed by a worked example.
The six trigonometric derivatives
Three rows without a sign, three rows with one. The reading rule fits in one word: the functions whose name starts with “co” carry the minus.
f(x)
Composite version
f′(x)
sinx
u′cosu
cosx
Example: (sin(3x))′=3cos(3x), which equals 3 at x=0, not 1: the factor 3 is half the mark.
cosx
−u′sinu
−sinx
Example: (cos(x2))′=−2xsin(x2), which is about −1.68 at x=1.
tanx
u′sec2u
sec2x=1+tan2x
Example: At x=4π, sec2x=2: the slope of the tangent line there is twice that of y=x.
cotx
−u′csc2u
−csc2x
Example: At x=4π, the slope is −2, exactly the opposite of that of tan at the same point.
secx
u′secutanu
secxtanx
Example: At x=3π, secx=2 and tanx=3, so the slope is 23≈3.46.
cscx
−u′cscucotu
−cscxcotx
Example: At x=3π, the slope is −32⋅31≈−0.667.
This whole table assumes the argument is in RADIANS. In degrees, (sinx)′=180πcosx≈0.01745cosx, which the calculator displays without warning when it was left in DEG mode.
Exponentials, logarithms, inverse and hyperbolic functions
Four families, one difficulty per row: the factor the base leaves behind, and the domain that is written down BEFORE differentiating.
f(x)
Domain
f′(x)
ex
everywhere
ex
Example: (e3x)′=3e3x: the 3 comes from u′, and it equals 3 at x=0.
ax, a>0
everywhere
axlna
Example: (5x)′=5xln5, so about 8.05 at x=1. The factor lna is the only trace of the base.
lnx
x>0
x1
Example: (ln(x2+1))′=x2+12x, which equals 1 at x=1: the u′ is in the numerator, never anywhere else.
logax
x>0
xlna1
Example: (log10x)′=xln101, that is about 0.0434 at x=10.
arcsinx
∣x∣<1 strictly
1−x21
Example: At x=0 the slope is 1; at x=0.8 it is already 0.61≈1.67, and it blows up at ±1.
arccosx
∣x∣<1 strictly
−1−x21
Example: Same absolute value as arcsin, opposite sign: the sum arcsinx+arccosx equals 2π for every x, so its derivative is zero.
arctanx
everywhere
1+x21
Example: At x=1, the slope is 0.5. No domain condition: it is the only inverse trigonometric function in this case.
sinhx and coshx
everywhere
coshx and sinhx
Example: No minus sign, unlike the circle: at x=1, sinh′(1)=cosh1≈1.543. They satisfy y′′=y, while sin and cos satisfy y′′=−y.
The domain is written before differentiating, not after: (ln(x−3))′=x−31 is wrong as long as x>3 has not been stated, and the test counts that line separately.
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.The factor u′ forgotten as soon as there are two layers
1 mark each time, and often 4 or 5 per test
What not to write
“(sin3x)′=cos3x”
What to write
“(sin3x)′=3cos3x: the table gives the outer derivative, the chain rule adds u′=3.”
At 0, sin3x rises three times as fast as sinx: its tangent line has slope 3cos0=3 (green). Without the factor u′=3, you find slope 1 (red), that of sinx.
Why: This is the most profitable mistake to fix in the whole chapter, because it repeats: a test that has it once has it everywhere. The check is mechanical: count the nested functions and count the factors.
2.Confusing cos(x2) and cos2x
the whole question, and textbooks often put both in the same sum
What not to write
“(cos2x)′=−2xsin(x2)”
What to write
“cos2x=(cosx)2: the square is OUTSIDE, so (cos2x)′=2cosx⋅(−sinx)=−sin(2x). It is cos(x2) that gives −2xsin(x2).”
cos2x (orange) stays between 0 and 1 and keeps a fixed period; cos(x2) (blue) oscillates faster and faster. Two different functions, two different derivatives.
Why: The notation cos2x hides the parentheses that say which function is on the outside. Rewrite it before differentiating: the outer function of (cosx)2 is the square, that of cos(x2) is the cosine.
3.Differentiating 2x with the power rule
the whole question
What not to write
“(2x)′=x2x−1”
What to write
“(2x)′=2xln2: the exponent is the VARIABLE, so this is not a power but an exponential.”
At x=1, the tangent line to y=2x has slope 2ln2≈1.39 (green). The power rule would give 1⋅20=1 (red): a line that cuts through the curve.
Why: The rule (xn)′=nxn−1 assumes the exponent is CONSTANT. Here it is the base that is constant: the two rules apply to different objects, and only the position of the x tells them apart.
4.Differentiating the logarithm before writing its domain
2 marks, plus the increasing/decreasing question that follows
What not to write
“f(x)=ln(x2−4) so f′(x)=x2−42x, and f′(0)=0.”
What to write
“f is defined only for x2−4>0, that is ∣x∣>2. The formula f′(x)=x2−42x holds only there, and f′(0) does not exist.”
ln(x2−4) exists only for ∣x∣>2: two branches, nothing between −2 and 2. The formula x2−42x can be computed at 0, but there is no curve at 0, so no slope.
Why: The formula for the derivative can still be computed where the function does not exist: it does not complain. This is the one chapter where the domain must be written BEFORE differentiating, and the marker looks for that line.
5.Believing that ln(5x) and lnx have different derivatives
1 mark
What not to write
“(ln5x)′=5x1”
What to write
“(ln5x)′=5x5=x1, the same derivative as lnx, which can also be seen from ln5x=ln5+lnx.”
ln5x=ln5+lnx: the curve is that of lnx raised by ln5. The tangent lines at x=1 are parallel, with slope 1: (ln5x)′=x1, not 5x1.
Why: A constant factor inside a logarithm becomes an ADDITIVE constant, so it vanishes when you differentiate. Simplifying the logarithm before differentiating avoids the mistake and saves time.
6.Applying a power rule to xx, where base AND exponent vary
the whole question
What not to write
“(xx)′=x⋅xx−1=xx”
What to write
“No rule applies. Set y=xx, so lny=xlnx, and differentiate: yy′=lnx+1, hence y′=xx(lnx+1).”
At x=2, the tangent line to y=xx has slope 4(ln2+1)≈6.77 (green). The power rule gives 4 (red): a line that is far too flat.
Why: The power rule freezes the exponent, the exponential rule freezes the base: when both move, only logarithmic differentiation is left. The wrong answer xx happens to be right at x=1, where both give 1, which is why a check at x=1 proves nothing; at x=e it gives ee instead of 2ee.
7.Stopping logarithmic differentiation at yy′
2 marks, half the question
What not to write
“yy′=x3+2x+11, so y′=x3+2x+11.”
What to write
“y′=y(x3+2x+11), where y is replaced by the original expression.”
Why: yy′ is the RELATIVE rate of change, not the derivative. The final multiplication is the step everyone skips because the interesting computation is already done.
8.The minus sign of arccos, and the extra one on cosh
1 mark each, and the increasing/decreasing behaviour of the whole analysis
What not to write
“(arccosx)′=1−x21” and “(coshx)′=−sinhx”
What to write
“(arccosx)′=1−x2−1, because arccos is DECREASING” and “(coshx)′=+sinhx, there is no minus sign for the hyperbolic functions.”
arcsin rises and arccos falls: that is all the minus sign is about. At both edges x=±1 the tangent lines are vertical, so the derivative does not exist there.
Why: The reflex “co means minus” comes from circular trigonometry and carries over neither to the inverse functions nor to the hyperbolic ones. The check is visual: a decreasing function has a negative derivative, an increasing function a positive one.
9.Simplifying x2 to x in the derivative of arccos(x3)
1 mark, and the wrong increasing/decreasing behaviour on the whole left half
What not to write
“(arccosx3)′=xx2−93”
What to write
“1−x29=∣x∣x2−9, so the derivative is ∣x∣x2−93, positive on the whole domain ∣x∣≥3.”
Why: x2=∣x∣, never x: the square root is positive. The wrong form gives −0.15 at x=−5, a negative slope for a function that increases, which the check by increasing/decreasing behaviour catches.
10.Confusing the absolute rate of change with the relative rate
the whole question, since the model is wrong from the first line
What not to write
“The population grows by 3% per year, so P′=0.03.”
What to write
“A percentage is a RELATIVE rate: PP′=0.03, so P′=0.03P and P(t)=P0e0.03t.”
Starting from P0=100: with P′=0.03P (green), the population more than quadruples in 50 years; with P′=0.03 (red), it gains 1.5 individuals. A percentage is a RELATIVE rate.
Why: P′=0.03 describes LINEAR growth of three hundredths of an individual per year. The word “percentage” in a problem always means PP′, and that is what produces the exponential.
Which method to choose
Which rule, depending on the FORM of the expression
Look at where the x is, in the base or in the exponent, before looking at which function it is. The position of the x decides, not the chapter.
Three neighbouring curves and three different rules: constant exponent, constant base, then both variable. It is the position of the x in the expression that decides, never the shape of the graph.
If variable base, CONSTANT exponent, like x5 or x−1/2 → power rule (xn)′=nxn−1
Example: (x7)′=7x6
If CONSTANT base, variable exponent, like 2x or e3x → exponential rule (au)′=aulna⋅u′
Example: (5x2)′=5x2ln5⋅2x
With a=e the factor lna equals 1, which explains why e is the working base.
If base AND exponent both variable, like xx or (sinx)x → logarithmic differentiation, no other rule exists
Example: y=xsinx, lny=sinxlnx
If product or quotient of at least THREE factors, with exponents or roots → logarithmic differentiation too, for convenience: the logarithm turns products into sums
Example: y=(x2+1)4x32x+1, taken apart below
The direct computation is possible, but it stacks two quotient rules and a product rule: that is where sign errors are born.
If a transcendental function applied to something other than x alone → the line of the table, multiplied by u′
Example: (ln(3x2+1))′=3x2+16x
If no branch applies, the expression is a sum: differentiate term by term and you land on one of the branches for each term.
Which exponential model, depending on the wording of the problem
The Part B problems all look alike. It is a precise phrase in the problem that points to the differential equation, and therefore to the model.
If “grows by k percent per year”, “doubles every T”, “half-life” → y′=ky, so y=y0ekt, with a constant RELATIVE rate
Example: half-life T gives k=Tln2, up to sign
If “tends to room temperature”, “approaches the value a” → Newton's law of cooling y′=k(y−a), so y=a+(y0−a)ekt
Example: a coffee at 80 degrees in a room at 20 degrees
It is the GAP from a that decays exponentially, not the temperature itself: setting u=y−a brings you back to the previous case.
If “carrying capacity”, “ceiling”, “the environment cannot feed more than K individuals” → logistic model y′=Kry(K−y), S-shaped curve
Example: maximum growth rate 4rK, reached when y=2K
If the problem gives two measurements and asks for the rate → do not guess the model, identify it: two points are enough to find y0 and k in y=y0ekt
Example: y(0)=1200 and y(10)=1620 give k=101ln12001620
In every case, write the differential equation BEFORE its solution: it carries the modelling marks, since solving it is taken for granted at this level.
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.
Logarithmic differentiation, written out for the marker
When to use it: Base and exponent both variable, or a product or quotient of at least three factors with exponents.
1State the function and announce the method: “Let y=(x2+1)4x32x+1, defined and strictly positive for x>0.”
2Take the logarithm of both sides, justifying that you are allowed to: “For x>0, y>0, so lny exists.” If the sign is not guaranteed, write ln∣y∣, whose derivative is the same.
3Break into a sum BEFORE differentiating; this is where the method pays off: “lny=3lnx+21ln(2x+1)−4ln(x2+1).”
4Differentiate both sides with respect to x, remembering that the left side gives yy′ and not y′.
5Multiply by y and replace y by its original expression: this is the step that concludes, and the one people skip.
Concluding sentence
For every x>0, y′=(x2+1)4x32x+1(x3+2x+11−x2+18x).
The trap: Writing lny=… then differentiating the left side as y1 instead of yy′. The y is a function of x, so the chain rule applies on this side of the equals sign too.
Marking: Usually 1 mark for the justified switch to the logarithm, 2 marks for the decomposition into a sum and its differentiation, 1 mark for the final multiplication by y.
Solving an exponential decay problem
When to use it: Half-life, carbon-14 dating, elimination of a drug, cooling: whenever a quantity decreases by a constant percentage.
1Name the function and its unit: “Let N(t) be the remaining mass, in grams, after t years.”
2Write the differential equation and justify it from the problem: “The decay is proportional to the quantity present, so N′=−kN with k>0.”
3Give the general solution N(t)=N0e−kt, then use the half-life data to find k: N(T)=2N0 gives e−kT=21, so k=Tln2.
4Answer the question asked by isolating the unknown with a logarithm, never by trial and error on the calculator.
5Check the order of magnitude: if the sample has lost a little less than half, the age must be a little less than one half-life.
Concluding sentence
With T=5730 years, k=5730ln2≈1.21×10−4 per year; since the sample contains 72% of the initial carbon 14, its age is t=k−ln0.72≈2716 years.
The trap: Forgetting the minus sign in N′=−kN, then finding a negative age and handing it in as is. A decreasing quantity has a negative derivative: the sign is decided before the computation, not after.
Marking: 1 mark for the differential equation stated and justified, 1 mark for finding k from the half-life, 1 mark for the numerical answer with its unit.
Check before you hand in
Five minutes of checking recover more marks than one more problem started in a hurry.
Have the calculator recompute the derivative at one point
On a TI, MATH then nDeriv(f(x),x,a) gives the numerical value of f′(a); on a Casio, the RUN menu and the dxd command. Compare with the value your expression gives at the same point.
For f(x)=sin3x at x=1, the calculator shows −2.97; an answer cos3x would give −0.99, exactly a third of that: this is the missing factor u′ showing up.
Count the factors before turning the page
Count the nested functions in the problem, then count the multiplied factors in your answer. The two numbers must match.
sin(e3x) contains three nested functions, so the derivative must contain three factors: cos(e3x)⋅e3x⋅3. Only two factors means a skipped line.
Compare the domain of the derivative with that of the function
The derivative cannot exist where the function does not exist. Reread the domain written at the start, and check that the formula you found is not used anywhere else.
x2−42x can be computed at x=0 even though ln(x2−4) is not defined there: the computation never complains, it is up to you to do so.
Check the sign against increasing/decreasing behaviour
An increasing function has a positive derivative on its domain, a decreasing function a negative one. Test the formula at a convenient point.
arccos decreases from π to 0, so a derivative found positive is wrong; e−x decreases, so −e−x is indeed negative, while +e−x betrays a lost sign.
Check RADIAN mode before any trigonometric derivative
Every formula in the chapter assumes the argument is in radians. A three-second check: sin(0.01) must display 0.00999…, not 0.000174.
In degree mode, (sinx)′ equals 180πcosx: a numerical check would then give 0.0175 times the expected value, which looks like a computation error and is not one.
The typical problem, taken apart
Differentiating a quotient of three factors, then evaluating the slope at x=1
Let y=(x2+1)4x32x+1 for x>0.
Compute y′ by logarithmic differentiation, then give the slope of the tangent line at the point with x-coordinate 1.
Step 1
For x>0, 2x+1>0 and x3>0, so y>0 and lny exists. Write lny=ln(x3)+ln(2x+1)−ln((x2+1)4).
Why
Justifying positivity before taking the logarithm is the first mark on the marking scheme: without it, writing lny makes no sense and the marker notes it.
Step 2
The exponents come down in front: lny=3lnx+21ln(2x+1)−4ln(x2+1).
Why
This is where the method saves time: the root becomes a half and the fourth power a coefficient. The direct computation would still have a quotient to differentiate.
Step 3
Differentiate both sides: yy′=x3+21⋅2x+12−4⋅x2+12x=x3+2x+11−x2+18x.
Why
The left side gives yy′ and not y1: y is a function of x, so the chain rule applies on both sides of the equals sign. Each term on the right is a uu′.
Step 4
Multiply by y: y′=(x2+1)4x32x+1(x3+2x+11−x2+18x).
Why
The step skipped on half of the tests. As long as you have not multiplied by y, you hold the RELATIVE rate, which is not the derivative asked for.
Step 5
At x=1: y(1)=241⋅3=163 and the parenthesis equals 3+31−4=−32, so y′(1)=163⋅(−32)=−243.
Why
Evaluate AFTER differentiating symbolically. Substituting x=1 as early as the third step would lose the general formula, which the next question almost always asks for.
Step 6
Check: −243≈−0.0722, and the calculator gives nDeriv≈−0.0722. The negative sign is consistent, since the parenthesis is negative and y is positive.
Why
A numerical check at one point catches both a sign error and a forgotten term, for the price of ten seconds on the calculator.
The conclusion, written out
For every x>0, y′=(x2+1)4x32x+1(x3+2x+11−x2+18x), and the tangent line at the point with x-coordinate 1 has slope −243≈−0.072.
The classic mistake on this problem: Writing lny=3lnx+21ln(2x+1)−4ln(x2+1) then concluding y′=x3+2x+11−x2+18x: the final multiplication by y is missing, and the result would be −32 instead of −243, that is about nine times too large.
Learn by heart
•The three trigonometric derivatives starting with “co” carry the MINUS sign, the other three do not.
•(ax)′=axlna, and (xn)′=nxn−1: it is the position of the x, base or exponent, that picks the rule.
•(lnu)′=uu′, but the DOMAIN u>0 is written before differentiating, not after.
•cos2x=(cosx)2, square outside; cos(x2), square inside. Rewrite before differentiating.
•As many factors in the derivative as nested functions in the problem.
•Logarithmic differentiation: ln, break into a sum, differentiate, THEN multiply by y.
•(arcsinx)′=1−x21 and (arccosx)′=−1−x21, for ∣x∣<1 strictly.
•(coshx)′=sinhx and (sinhx)′=coshx: no minus sign for the hyperbolic functions.
•A percentage in a problem means yy′, the RELATIVE rate, hence an exponential.
Frequently asked questions
How do you differentiate a function like two to the power x?
You do not use the power rule: the exponent is the variable, not the base. The derivative of two to the power x is two to the power x multiplied by the natural logarithm of two. More generally, the derivative of a to the power u is a to the power u, times the natural logarithm of a, times the derivative of u. With base e, this factor equals one and disappears.
When should you use logarithmic differentiation?
In two cases. First, when the base and the exponent both contain the variable, as in x to the power x: no other rule exists then. Second, for convenience, when the expression is a product or a quotient of at least three factors with roots or powers: the logarithm turns these products into sums, which are much easier to differentiate.
Why is the derivative of arccosine negative?
Because the arccosine function decreases: it starts at pi when x equals minus one and goes down to zero when x equals one. A decreasing function has a negative derivative everywhere, hence the minus sign in front of one over the square root of one minus x squared. Arcsine, on the other hand, increases, so its derivative is positive.
How do you find the domain of the derivative of a logarithm?
First write the domain of the function, by requiring the argument of the logarithm to be strictly positive, then differentiate. The formula you get can often still be computed outside that domain, but it has no meaning there. This is the most costly trap of the chapter: the domain is written before the derivative, never after.
What does a growth rate of three percent per year mean?
It means that the derivative divided by the quantity equals zero point zero three, not that the derivative equals zero point zero three. It is a relative rate, which leads to exponential growth of the form y equals y zero times e to the power zero point zero three t. An absolute rate, on the other hand, would give a straight line.
Practise it
Corrected exercises: Derivatives of trigonometric, exponential and logarithmic functions, 201-SN2-RE
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 work on transcendental derivatives at the level of real CEGEP tests, from the trigonometric table all the way to logarithmic differentiation and exponential models.