3.3 Differentiation Rules

The constant and power rules, the sum, difference and constant multiple rules, the product rule with the area proof that explains its two terms, the quotient rule and why its order matters, and the discipline of deciding which rule applies outermost when several are needed at once.

Subject: Calculus I · 70 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. Differentiation Rules

Title

Calculus I · Section 3.3

Derivatives of every polynomial and rational function, without a single limit

2. What this lesson gives you

Objectives

In Section 3.2 every derivative cost a limit calculation. This lesson proves a handful of rules once, from the definition, and then uses them in place of the definition from now on.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 216 — learning objectives 3.3.1 to 3.3.6

In the last section you found derivatives the honest way: write the difference quotient, expand, cancel, divide by h, and let h shrink. That is the definition, and it never stops being the definition. But doing it for every function you meet would be like proving the multiplication table each time you needed seven times eight.

This lesson proves a small set of rules once, carefully, from the definition. Two of them handle the building blocks, constants and powers of x. The others handle the ways building blocks are combined: adding, scaling, multiplying and dividing. Together they reach every polynomial and every rational function.

The last parts put the rules to work: tangent lines, points where the tangent is flat, and the velocity of a moving object. Keep an eye on the checks at the end of each example; a derivative you can check is a derivative you can trust.

3. Before anything new: the definition

Warm-up

Discussion prompt

Write the limit definition of the derivative of f at x, then use it to find the derivative of x squared. Keep your work; this lesson is about never having to do it again.

Write your answer before revealing it. The definition is worth having word-perfect, because every rule in this lesson is proved from it, and you will want to follow those proofs rather than take them on trust.

Notice the shape of the calculation for x squared. The expansion produces a term with no h, which cancels against the subtracted x squared; a term with exactly one factor of h, which survives the division by h; and a term with h squared, which is still carrying a factor of h after the division and so disappears in the limit.

That three-way split, cancel, survive, vanish, is the whole story of the power rule. Watch for it again in Example 3.18 and in the proof of Theorem 3.3.

4. Why rules instead of limits

Intuition

The definition always works, but it gets expensive fast. The derivative of the square root needed a conjugate trick; the cube root needs a harder one, and a quotient of two polynomials is worse still.

\[ \frac{d}{dx}\left(\sqrt{x}\right) = \frac{1}{2\sqrt{x}} \qquad \frac{d}{dx}\left(\sqrt[3]{x}\right) = \;? \]

The plan: prove a rule for the building blocks (constants and powers), then a rule for each way of combining them (sums, multiples, products, quotients). Every polynomial and rational function is built that way, so its derivative follows mechanically.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 216 — introduction and The Basic Rules

The derivative of the square root came out of a conjugate trick: multiply top and bottom by the sum of two roots so that the roots in the numerator cancel. It works, but you had to know the trick. For the cube root you would need a cleverer identity, and for a messy quotient of polynomials the algebra would fill a page.

The way out is structural. Complicated functions are built from simple ones by a few operations. If you know the derivative of each simple piece, and you know how each operation affects derivatives, you can differentiate anything built that way without ever writing another limit.

That is the plan for the whole lesson: building blocks first, then one rule per operation.

5. Constants and powers

Section

Part 1

6. The constant rule

Concept

Figure (svg): The horizontal line y equals 8, with two points on it at x and x plus h joined by a flat secant; the rise between them is zero.

Whatever two points you pick, the rise is zero, so every difference quotient is zero and so is its limit.

Put a constant function into the definition. Both function values are the same number c, so the numerator is zero before the limit is even taken.

\[ f'(x) = \lim_{h\to 0}\frac{c - c}{h} = \lim_{h\to 0} 0 = 0 \]

Constant rule (Theorem 3.2) — The derivative of a constant function is zero: a horizontal line has slope zero everywhere.

\[ \frac{d}{dx}(c) = 0 \]

\[ \text{Example 3.17: } f(x) = 8 \;\Rightarrow\; f'(x) = 0 \]

\[ \text{Checkpoint 3.11: } g(x) = -3 \;\Rightarrow\; g'(x) = 0 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 216-217 — Theorem 3.2, Example 3.17, Checkpoint 3.11

Start with the simplest function there is. A constant function has the same value everywhere, so its graph is a horizontal line, and a horizontal line has slope zero. The figure makes the same point with numbers: any two points on the line have the same height, so the rise between them is zero.

In the definition this shows up before any limit is needed. The numerator is c minus c, which is zero for every h, so the difference quotient is zero and so is its limit. Example 3.17 and Checkpoint 3.11 are one-line uses of the rule: the derivative of 8 is zero, and so is the derivative of minus 3.

The rule matters more than it looks. Every polynomial ends in a constant term, and the constant rule is what lets that term drop out of the derivative. Shifting a graph up or down never changes its slopes.

7. Example 3.18: x cubed from the definition

Worked example

Find the derivative of x cubed straight from the limit definition, watching which terms survive.

\[ \frac{d}{dx}\left(x^3\right) = \lim_{h\to 0}\frac{(x+h)^3 - x^3}{h} \]

Expand the cube

Why: The first term is x cubed and the second is three x squared h; every other term has at least two factors of h.

\[ (x+h)^3 = x^3 + 3x^2h + 3xh^2 + h^3 \]

Cancel the x cubed terms

Why: Only terms containing h remain in the numerator.

\[ \lim_{h\to 0}\frac{3x^2h + 3xh^2 + h^3}{h} \]

Factor out h

Why: Every surviving term has a factor of h.

\[ \lim_{h\to 0}\frac{h\left(3x^2 + 3xh + h^2\right)}{h} \]

Cancel h

Why: Legal because h is not zero inside the limit.

\[ \lim_{h\to 0}\left(3x^2 + 3xh + h^2\right) \]

Let h go to zero

Why: Only the term without h is left.

\[ \frac{d}{dx}\left(x^3\right) = 3x^2 \]

Check with a small h

Why: At x equal to 2 the answer predicts slope 12; a difference quotient with h equal to one thousandth agrees to within the leftover three x h.

\[ \frac{(2.001)^3 - 2^3}{0.001} = 12.006001 \approx 12 = 3(2)^2 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 217-218 — Example 3.18

This is the last derivative of a power you will do by the definition, and it is worth doing slowly because the general proof follows exactly the same steps.

The expansion is the only real work. When you multiply out x plus h, cubed, you get one term with no h at all, three terms that each contain exactly one h and add up to three x squared h, and the remaining terms, which contain h squared or h cubed. The x cubed cancels. Dividing by h leaves three x squared plus terms that still contain h, and those vanish in the limit.

The check shows you the leftover terms at work. With h equal to one thousandth, the difference quotient at 2 is 12.006001, not exactly 12. The extra 0.006001 is precisely three x h plus h squared at x equal to 2, the terms the limit removes.

8. Where three x squared h comes from

Picture it

Figure (svg): A cube of side x drawn in perspective, with three coloured slabs of thickness h attached to its right face, its top face and its front face; the thin edge pieces where two slabs would meet are left empty.

Almost all of the new volume sits in the three slabs. The pieces left out are thinner by a factor of h and vanish first as h shrinks.

Picture x cubed as the volume of a cube of side x. Lengthen every side by h and the new volume is made of three slabs of volume x squared times h, plus thin edge pieces and a tiny corner.

\[ (x+h)^3 - x^3 = \underbrace{3x^2h}_{\text{slabs}} + \underbrace{3xh^2 + h^3}_{\text{edges, corner}} \]

Here is a way to see the algebra. Think of x cubed as the volume of a cube whose sides have length x. Now lengthen every side by a small amount h. The extra volume comes mostly from three thin slabs glued onto three faces of the cube, each one x by x by h.

What is missing from the picture are the thin edge pieces where two slabs would meet, each x by h by h, and the tiny corner cube, h by h by h. Those are the terms three x h squared and h cubed. They are thinner than the slabs by a whole factor of h.

When you divide the extra volume by h, the slabs give three x squared, and the edges and corner still carry a factor of h, so they vanish as h shrinks. The coefficient 3 in the derivative is literally counting the three slabs; in dimension n there would be n of them.

9. Shrink h and watch the difference quotient

Tweak it

Parameter explorer

The curve is the difference quotient of x cubed, drawn as a function of x for the chosen h. Slide h toward zero. Which familiar curve does it settle onto?

\[ \frac{(x+{h})^3 - x^3}{{h}} = 3x^2 + 3({h})x + ({h})^2 \]

  • h — from 0.05 to 2: step h

Each position of the slider fixes a step size h and draws the difference quotient of x cubed as a function of x. The readout shows the algebra: it equals three x squared, plus three h times x, plus h squared.

Start with h large and the curve is a parabola shifted to the left and lifted up. As you slide h toward zero, both extra terms shrink, and the curve settles onto the parabola three x squared, which is the derivative. The limit in the definition is exactly this settling.

Notice that the settling happens at every x at once. That is why the derivative is a new function and not just a number: taking the limit at each x separately gives a whole curve of slopes.

10. Checkpoint 3.12: guess the derivative of x to the fourth

Prediction

\[ \frac{d}{dx}\left(x^2\right) = 2x, \qquad \frac{d}{dx}\left(x^3\right) = 3x^2, \qquad \frac{d}{dx}\left(x^4\right) = \;? \]

Predict first

Follow the pattern. What is the derivative of x⁴?

  • 4x³
  • 4x⁴
  • 3x⁴
  • x³

Correct: 4x³

Why: The exponent comes down as the coefficient and the exponent drops by one. The definition confirms it: expanding (x + h)⁴ gives x⁴ + 4x³h + terms with h², so the difference quotient tends to 4x³.

\[ (x+h)^4 - x^4 = 4x^3h + 6x^2h^2 + 4xh^3 + h^4 \quad\Rightarrow\quad \frac{d}{dx}\left(x^4\right) = 4x^3 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 218 — Checkpoint 3.12

Commit to an answer before revealing. Two data points make a pattern: the derivative of x squared is two x, and of x cubed is three x squared. In each case the exponent came down in front and the new exponent is one less.

Guessing from a pattern is not a proof, so the slide backs it up. Expanding x plus h to the fourth gives x to the fourth, then four x cubed h, then terms with at least two factors of h. The same cancel, survive, vanish split leaves four x cubed.

If you chose four x to the fourth, you brought the exponent down but forgot to lower it. A quick degree check catches that: the derivative of a degree-four polynomial has degree three.

11. The power rule (Theorem 3.3)

Concept

The cube argument works for any positive whole number n. The binomial theorem says which terms of the expansion carry exactly one factor of h.

\[ (x+h)^n = x^n + nx^{n-1}h + \binom{n}{2}x^{n-2}h^2 + \cdots + h^n \]

\[ \frac{(x+h)^n - x^n}{h} = nx^{n-1} + \binom{n}{2}x^{n-2}h + \cdots + h^{n-1} \]

\[ f'(x) = \lim_{h\to 0}\left(nx^{n-1} + \binom{n}{2}x^{n-2}h + \cdots + h^{n-1}\right) = nx^{n-1} \]

Power rule — For a positive integer n, the derivative of x to the n is n times x to the n minus 1. The rule is later extended to negative, rational and all real exponents.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 218-219 — Theorem 3.3 and its proof

The proof is the cube argument in general form. The binomial theorem expands x plus h to the n: first x to the n, then n x to the n minus 1 times h, then terms carrying h squared, h cubed and so on. The binomial coefficient n choose 2 in the third term does not matter; only the fact that it multiplies h squared does.

Subtract x to the n, divide by h, and every term except n x to the n minus 1 still contains at least one factor of h. Let h go to zero and those terms disappear.

The theorem is stated for positive whole numbers n, because that is what the binomial theorem gives. Later in this lesson the quotient rule extends it to negative whole numbers, and later in the chapter it is extended to every real exponent. The formula never changes.

12. Reading the power rule piece by piece

Notation

Annotate

On: \( \frac{d}{dx}\left(x^{n}\right) = n\,x^{n-1} \)

  • Leibniz notation for 'the derivative with respect to x of'. It is an instruction, not a fraction you can cancel.
  • The base is the variable and the exponent is a fixed number. That is what makes it a power function.
  • The old exponent comes down as a coefficient: it counts the n slabs in the picture.
  • The new exponent is one less. A power of degree n has a slope of degree n − 1.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 218 — Theorem 3.3

Step through the annotations one at a time. The d over d x in front is an operator: it takes a function and returns its derivative. It looks like a fraction, and later you will see why that is a useful way to think, but for now read it as an instruction.

The most important annotation is the second. The rule is about a variable base raised to a fixed exponent. That condition is exactly what fails in the trap a few slides from now.

The coefficient n and the new exponent n minus 1 go together. A reliable self-check is to differentiate a power in your head and then confirm the degree dropped by exactly one and the old exponent is sitting in front.

13. Example 3.19 and Checkpoint 3.13: x to the tenth and x to the seventh

Worked example

\[ \text{(a) } f(x) = x^{10} \qquad \text{(b) } f(x) = x^{7} \]

Identify n in (a)

Why: The exponent is ten.

\[ n = 10 \]

Apply the power rule to (a)

Why: Bring ten down and reduce the exponent by one.

\[ f'(x) = 10x^{10-1} = 10x^9 \]

Apply it to (b), Checkpoint 3.13

Why: Now n is seven.

\[ f'(x) = 7x^{7-1} = 7x^6 \]

Check (a) with a difference quotient at x equal to 1

Why: The rule predicts slope 10 there; a small step gives almost exactly that.

\[ \frac{(1.001)^{10} - 1}{0.001} \approx 10.045 \approx 10 = 10(1)^9 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 219 — Example 3.19 and Checkpoint 3.13

Compare this with the page of algebra that the definition would need for x to the tenth: an expansion with eleven terms. The rule does it in one line.

The only thing to get right is the bookkeeping: identify n, bring it down, subtract one from it. For x to the tenth that gives ten x to the ninth, and for Checkpoint 3.13, x to the seventh, it gives seven x to the sixth.

The check uses a small step at x equal to 1. The rule says the slope there is 10, and the difference quotient with h equal to one thousandth is about 10.045. The small excess is the leftover terms again, which is exactly what you should expect when h is small but not zero.

14. How steep is x to the tenth at x = 1.1?

Estimation

\[ f(x) = x^{10}, \qquad f'(1.1) = 10(1.1)^9 \]

Predict first

Estimate before computing: roughly how steep is the graph of x¹⁰ at x = 1.1?

  • About 1
  • About 10
  • About 24
  • About 100

Correct: About 24

Why: 1.1⁹ is about 2.36, so the slope is 10 × 2.36 ≈ 23.6. Ten percent past x = 1, the slope has more than doubled from 10: high powers steepen very quickly, which is why their graphs look like walls.

\[ f'(1.1) = 10(1.1)^9 \approx 10(2.358) \approx 23.6 \]

Make your estimate before revealing. Most people guess close to 10, because the slope at 1 was 10 and 1.1 is only a little bigger than 1.

But the derivative is ten times 1.1 to the ninth, and 1.1 to the ninth is about 2.36. So the slope is about 23.6, more than double the slope at 1. High powers grow and steepen very fast once x passes 1; that is why the graph of x to the tenth looks almost flat near zero and then shoots up like a wall.

The lesson for estimating derivatives is that the derivative is itself a function with its own growth. A small change in x can mean a large change in slope.

15. Trap: the power rule on a variable exponent

Trap

The trap

The power rule gets applied to anything with an exponent:

\[ \frac{d}{dx}\left(3^{x}\right) = x\cdot 3^{x-1} ? \]

Wrong. Here the variable is the exponent, not the base.

The fix

The power rule needs a variable base and a fixed exponent. In three to the x the base is fixed and the exponent varies: an exponential function, with its own rule in Section 3.9.

Figure (svg): The curve y equals 3 to the x near x equals 2, with its true tangent of slope about 9.89 at the point (2, 9) and a flatter dashed line of slope 6, the value the misapplied power rule would give.

The misapplied rule gives a line that is visibly not tangent.

\[ \text{at } x = 2: \quad 2\cdot 3^{1} = 6 \;\ne\; 9.89 \]

This error comes from matching the shape of the notation instead of the meaning. Three to the x has an exponent, so the power rule looks applicable. But in a power function the base is the variable and the exponent is fixed; here it is the other way round.

The picture settles it. At the point where x equals 2, the misapplied rule predicts a slope of 6, and the dashed line with that slope is clearly not tangent: it is too flat. The true slope is about 9.89, which comes from a different rule involving the natural logarithm of 3.

Before using the power rule, ask one question: is the variable in the base, with a constant exponent? If not, the power rule does not apply. Exponential functions get their own rule later in the chapter.

16. Sums, differences and multiples

Section

Part 2

17. Theorem 3.4: sums, differences, constant multiples

Concept

\[ (f+g)' = f' + g', \qquad (f-g)' = f' - g', \qquad (kf)' = kf' \]

The proof of the sum rule is one regrouping. Put the sum into the definition and separate the f terms from the g terms.

\[ j(x) = f(x) + g(x) \]

\[ j'(x) = \lim_{h\to 0}\frac{\left(f(x+h) + g(x+h)\right) - \left(f(x) + g(x)\right)}{h} \]

\[ j'(x) = \lim_{h\to 0}\left(\frac{f(x+h) - f(x)}{h} + \frac{g(x+h) - g(x)}{h}\right) \]

\[ j'(x) = f'(x) + g'(x) \]

The last step is the sum law for limits. The difference and constant multiple rules follow the same way.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 219-220 — Theorem 3.4 and proof of the sum rule

These three rules say that differentiation respects addition, subtraction and scaling by constants. That sounds obvious, but it is a fact about limits, and the proof shows exactly where it comes from.

Follow the lines of the sum rule proof. The numerator of the difference quotient of f plus g contains four function values. Regroup them so the two f values sit together and the two g values sit together, and the single fraction splits into two difference quotients. The sum law for limits then says the limit of the sum is the sum of the limits, which is f prime plus g prime.

Together with the power and constant rules, these give you the derivative of every polynomial: split it into terms, pull each coefficient out, and differentiate each power.

18. Scaling a graph scales every slope

Picture it

Figure (svg): Left: the parabolas y equals x squared and y equals 3 x squared for x from minus one half to 2. Right: their derivatives, the lines 2x and 6x; at every x the second line is three times as high as the first.

Stretching a graph vertically by 3 makes every rise three times as big over the same run, so every slope triples.

Multiplying x squared by 3 stretches the graph vertically. Over any run, every rise is three times as large, so the derivative line on the right is three times as steep everywhere.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 220 — Figure 3.18

The left panel shows x squared and three x squared. The second is the first stretched vertically by a factor of 3: every height is tripled.

Now think about slope as rise over run. The run between two x values is unchanged by a vertical stretch, but the rise is tripled, so every secant slope and every tangent slope is tripled. The right panel shows the result: the derivative of three x squared is the line six x, three times as steep as the line two x at every x.

This is the constant multiple rule as a picture, and it is Figure 3.18 in the book. Keep it in mind whenever you pull a constant out in front of a derivative.

19. Example 3.20: three times x squared

Worked example

Find the derivative of g and compare it with the derivative of f.

\[ g(x) = 3x^2, \qquad f(x) = x^2 \]

Pull the constant out

Why: Constant multiple rule.

\[ g'(x) = \frac{d}{dx}\left(3x^2\right) = 3\frac{d}{dx}\left(x^2\right) \]

Apply the power rule

Why: The derivative of x squared is two x.

\[ g'(x) = 3(2x) \]

Simplify

Why: Multiply the constants.

\[ g'(x) = 6x \]

Compare with f prime

Why: The derivative of f is two x, so g prime is three times f prime at every x.

\[ g'(x) = 6x = 3(2x) = 3f'(x) \]

Check at x equal to 1 with a small step

Why: The rule predicts 6.

\[ \frac{3(1.001)^2 - 3(1)^2}{0.001} = 6.003 \approx 6 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 220 — Example 3.20

The constant multiple rule lets you set the 3 aside, differentiate x squared, and multiply back. That is two moves, and the slide keeps them as separate lines so you can see each rule acting.

The comparison step is the point of the example. The derivative of x squared is two x, and the derivative of three x squared is six x, which is three times two x at every single value of x, exactly what the stretched graphs showed.

The numerical check at x equal to 1 gives 6.003 for a small step, close to the predicted 6. As with every check in this lesson, the small excess is the price of using a small h instead of the limit.

20. Example 3.21 and Checkpoint 3.14: polynomials termwise

Worked example

\[ \text{(a) } f(x) = 2x^5 + 7 \qquad \text{(b) } f(x) = 2x^3 - 6x^2 + 3 \]

Split (a) with the sum rule

Why: Differentiate each term separately.

\[ f'(x) = \frac{d}{dx}\left(2x^5\right) + \frac{d}{dx}(7) \]

Constant multiple rule

Why: Pull the 2 out of the first term.

\[ f'(x) = 2\frac{d}{dx}\left(x^5\right) + \frac{d}{dx}(7) \]

Power rule and constant rule

Why: Five x to the fourth; the constant contributes zero.

\[ f'(x) = 2\left(5x^4\right) + 0 \]

Simplify

Why: Multiply the constants.

\[ f'(x) = 10x^4 \]

Differentiate (b) term by term

Why: The same three rules, applied to each of the three terms at once.

\[ f'(x) = 2\left(3x^2\right) - 6(2x) + 0 = 6x^2 - 12x \]

Check both at x equal to 1

Why: Answer (a) predicts 10 and answer (b) predicts minus 6; small-step difference quotients agree.

\[ \frac{f_a(1.001) - f_a(1)}{0.001} \approx 10.02, \qquad \frac{f_b(1.001) - f_b(1)}{0.001} \approx -6.000 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 220-221 — Example 3.21 and Checkpoint 3.14

The book writes this solution in Leibniz notation on purpose, so that you can see the order in which the rules act. The sum rule splits the function into two terms. The constant multiple rule pulls the 2 out. Then the power rule and the constant rule finish the job.

With practice all of that collapses into one line, and part (b), which is Checkpoint 3.14, shows the compressed form: each term is differentiated at once, the constant term disappears, and you get six x squared minus twelve x.

Notice that the 7 in part (a) and the 3 in part (b) contribute nothing. Adding a constant shifts the graph up without tilting it, so it cannot change any slope. The checks at x equal to 1 confirm both answers.

21. Example 3.22: the tangent line at x = 1

Worked example

Find an equation of the line tangent to the graph of f at x equal to 1.

\[ f(x) = x^2 - 4x + 6 \]

Find the point

Why: Evaluate f at 1.

\[ f(1) = 1^2 - 4(1) + 6 = 3 \quad\Rightarrow\quad (1, 3) \]

Differentiate

Why: Termwise, with the power and constant rules.

\[ f'(x) = 2x - 4 \]

Find the slope at x equal to 1

Why: Evaluate the derivative at the point, not before.

\[ f'(1) = 2(1) - 4 = -2 \]

Write the point-slope form

Why: Point (1, 3), slope minus 2.

\[ y - 3 = -2(x - 1) \]

Rearrange to slope-intercept form

Why: Expand and add 3.

\[ y = -2x + 5 \]

Figure (svg): The parabola y equals x squared minus 4x plus 6 with its vertex at (2, 2), and the tangent line y equals minus 2x plus 5 touching it at the point (1, 3).

The line touches the parabola at one point and runs along it there with slope minus 2.

Check that the line only touches

Why: Setting the parabola equal to the line leaves a perfect square: one repeated meeting point, which is what tangency looks like for a parabola.

\[ x^2 - 4x + 6 = -2x + 5 \iff x^2 - 2x + 1 = 0 \iff (x-1)^2 = 0 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 221 — Example 3.22

A line needs a point and a slope. The point comes from the function itself: evaluate f at 1 and you get 3. The slope comes from the derivative: differentiate first, then evaluate at 1, giving minus 2.

Point-slope form writes the line straight from those two numbers, and rearranging gives slope-intercept form, y equals minus two x plus five. In the figure the dashed line touches the parabola at the point one, three, and runs along it there.

The check is a nice one for parabolas. If you set the parabola equal to the line, you get a quadratic equation, and for a tangent line that equation has one repeated root. Here it is x minus 1, squared, equal to zero: the line meets the curve at x equal to 1 and nowhere else.

22. Checkpoint 3.15: point-slope form

Fill the middle

\[ f(x) = 3x^2 - 11 \quad \text{at } x = 2: \quad f(2) = 1, \quad f'(x) = 6x \]

Fill in the blanks

y - 1 = 12(x - 2) \quad\Longrightarrow\quad y = 12x - 23

Why: The slope is f′(2) = 6 · 2 = 12 and the point is (2, 1). Expanding y − 1 = 12(x − 2) gives y = 12x − 24 + 1 = 12x − 23.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 221 — Checkpoint 3.15

Fill in both blanks before checking. The slide gives you the ingredients: the point is two, one, because f of 2 is 12 minus 11, and the derivative is six x.

The slope of the tangent line is the derivative evaluated at 2, which is 12. Put the point and the slope into point-slope form, then expand: y minus 1 equals twelve x minus 24, so y equals twelve x minus 23.

If you got a different constant, check the sign when you moved the 1 across. That step is where most slips happen in tangent-line problems.

23. Trap: leaving x in the slope

Trap

The trap

A tangent line at x equal to 1, written with the derivative still unevaluated:

\[ y - 3 = (2x - 4)(x - 1) \]

Wrong. That is a parabola, not a line.

The fix

The slope of a tangent line is a number: the derivative evaluated at the point. Substitute first, then write the line.

\[ m = f'(1) = -2 \]

\[ y - 3 = -2(x - 1) \]

This mistake happens when the derivative is written down as a formula and then plugged into point-slope form before it has been evaluated. The result has x times x in it, so it is a parabola, and a tangent line cannot be a parabola.

The fix is an order of operations. The derivative is a function of x; the slope of the tangent line at one particular point is a single number, the value of that function at the point. Evaluate first, then write the line. A quick test is to look at your answer: if it has an x squared in it, it is not a line.

24. The product rule

Section

Part 3

25. The derivative of a product is not the product of derivatives

Concept

Test the tempting guess on a product you can already differentiate. Write x squared as x times x.

\[ \frac{d}{dx}\left(x^2\right) = 2x \qquad\text{but}\qquad \frac{d}{dx}(x)\cdot\frac{d}{dx}(x) = 1\cdot 1 = 1 \]

A second test, with x squared times x cubed, which is x to the fifth:

\[ \frac{d}{dx}\left(x^5\right) = 5x^4 \qquad\text{but}\qquad (2x)\left(3x^2\right) = 6x^3 \]

The guess gets the wrong coefficient and the wrong degree. Something else is needed.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 221 — The Product Rule, introduction

The sum rule worked so neatly that it is natural to hope products behave the same way. The quickest way to test a hope like that is on an example where you already know the answer.

Write x squared as x times x. The true derivative is two x, but the product of the derivatives is 1 times 1, which is 1. Try x squared times x cubed, which is x to the fifth: the true derivative is five x to the fourth, but the product of the derivatives is six x cubed.

Look at what went wrong in the second test. The guess is not just off by a constant; it has the wrong degree. That tells you the true rule must involve the original functions as well as their derivatives, not only the derivatives.

26. A growing rectangle

Picture it

Figure (svg): A rectangle of width f of x and height g of x. When x changes by h, the width grows by delta f and the height by delta g, adding a tall strip on the right with area delta f times g, a wide strip on top with area f times delta g, and a small corner of area delta f times delta g.

Two strips, one for each factor that moved, plus a corner that is small times small.

Think of the product as the area of a rectangle with sides f and g. Nudge x, and both sides change. The area gains a strip for the change in f, a strip for the change in g, and a small corner.

\[ \Delta(fg) = \Delta f\cdot g + f\cdot\Delta g + \Delta f\,\Delta g \]

Here is the picture that explains the product rule. Draw the product f times g as the area of a rectangle, with width f of x and height g of x. When x changes a little, the width changes by a small amount, delta f, and the height changes by a small amount, delta g.

The new rectangle contains the old one plus three new pieces. The tall strip on the right has width delta f and height g, so its area is delta f times g. The wide strip on top has width f and height delta g. And the small corner has area delta f times delta g.

The two strips are where the two terms of the product rule come from: one for each factor that changed. The corner is the only piece left to worry about, and the next slide shows why it does not matter.

27. Why the corner vanishes

Intuition

Divide the change in area by h and look at each piece.

\[ \frac{\Delta(fg)}{h} = \frac{\Delta f}{h}\,g + f\,\frac{\Delta g}{h} + \frac{\Delta f}{h}\,\Delta g \]

The first two pieces become f prime times g and f times g prime. The corner piece is a difference quotient times the change in g, and the change in g shrinks to zero with h, so the corner contributes nothing in the limit.

\[ \frac{\Delta f}{h}\,\Delta g \;\to\; f'(x)\cdot 0 = 0 \]

Divide the change in area by h, just as the definition of the derivative does. The strip for f becomes the difference quotient of f times g, which tends to f prime times g. The strip for g becomes f times the difference quotient of g, which tends to f times g prime.

The corner divided by h is the difference quotient of f times delta g. The difference quotient tends to f prime, a finite number, but delta g tends to zero, because g is continuous. A finite number times something going to zero goes to zero.

So the corner is small in a stronger sense than the strips: it is small times small, and even after one division by h it is still small. That is the whole reason the product rule has exactly two terms.

28. Theorem 3.5: the product rule

Concept

\[ \frac{d}{dx}\left(f(x)g(x)\right) = f'(x)g(x) + g'(x)f(x) \]

Product rule — The derivative of a product is the derivative of the first times the second, plus the derivative of the second times the first.

Two terms, one for each factor that can change. Each term differentiates exactly one factor and leaves the other alone.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 221-222 — Theorem 3.5

Here is the rule the picture predicted. Each of the two terms differentiates exactly one factor and leaves the other one alone. Say it as a sentence: derivative of the first times the second, plus derivative of the second times the first.

Because the two terms are added, their order does not matter. Many people write f times g prime plus f prime times g, and that is the same thing. The quotient rule, coming next, will not be so forgiving.

The rule extends naturally. With three factors you get three terms, one for each factor that is differentiated, and you will see that in Example 3.29.

29. Reading the product rule piece by piece

Notation

Annotate

On: \( j'(x) = f'(x)\,g(x) + g'(x)\,f(x) \)

  • The function being differentiated is a product: its last operation is multiplication.
  • The right-hand strip: f changes while g is held at its current value.
  • The top strip: g changes while f is held fixed.
  • A plus sign, so the order of the two terms does not matter. Contrast the quotient rule.

Step through each annotation and connect it to the rectangle. The first term is the tall strip: f changes while g stays at its current value. The second term is the wide strip: g changes while f stays fixed.

The plus sign is worth a note on its own. It means the product rule is symmetric, so you can write the terms in either order. That symmetry is lost in the quotient rule, which has a minus sign where this has a plus, and that is the main reason quotient answers come out with the wrong sign.

30. The proof: add and subtract one term

Concept

Start from the definition. The trick is to add and subtract f of x times g of x plus h, so that each factor changes one at a time.

\[ j'(x) = \lim_{h\to 0}\frac{f(x+h)g(x+h) - f(x)g(x)}{h} \]

\[ = \lim_{h\to 0}\frac{f(x+h)g(x+h) - f(x)g(x+h) + f(x)g(x+h) - f(x)g(x)}{h} \]

\[ = \lim_{h\to 0}\left(\frac{f(x+h) - f(x)}{h}\cdot g(x+h)\right) + \lim_{h\to 0}\left(\frac{g(x+h) - g(x)}{h}\cdot f(x)\right) \]

\[ = f'(x)g(x) + g'(x)f(x) \]

The last step uses continuity of g: a differentiable function is continuous, so g of x plus h tends to g of x.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 222 — proof of Theorem 3.5

The proof mirrors the rectangle. From the definition, the numerator is the new area minus the old area. The trick is to add and subtract the same quantity, f of x times g of x plus h, which is the area when only g has changed. Adding and subtracting it changes nothing, but it lets you change one factor at a time.

Now the numerator splits into two differences. The first has f changing with g held at x plus h; the second has g changing with f held at x. Each one divided by h is a difference quotient times a function value.

The last step needs care. As h goes to zero, g of x plus h goes to g of x, because a differentiable function is continuous. Without that fact, the first term would not settle down. With it, the two limits give f prime g plus g prime f.

31. Example 3.23: the product rule at a point

Worked example

Find j prime of 2 for j equal to f times g, knowing only these four numbers.

\[ f(2) = 3, \quad f'(2) = -4, \quad g(2) = 1, \quad g'(2) = 6 \]

Write the product rule at x equal to 2

Why: Only values at 2 are needed.

\[ j'(2) = f'(2)g(2) + g'(2)f(2) \]

Substitute

Why: Four numbers, in the right places.

\[ j'(2) = (-4)(1) + (6)(3) \]

Evaluate

Why: Minus 4 plus 18.

\[ j'(2) = 14 \]

Check with concrete functions

Why: Lines with these values and slopes at 2 are one choice of f and g; with u for x minus 2, their product expands and its u coefficient is the slope at 2.

\[ (3 - 4u)(1 + 6u) = 3 + 14u - 24u^2 \quad\Rightarrow\quad j'(2) = 14 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 222 — Example 3.23

This example shows something important: to use the product rule at a point, you do not need formulas for f and g at all. You need just four numbers, the value and the derivative of each function at that point.

Substitute carefully. The derivative of f at 2 multiplies the value of g at 2, and the derivative of g at 2 multiplies the value of f at 2. That gives minus 4 times 1, plus 6 times 3, which is 14.

The check invents functions that fit the data. The simplest are lines: f starting at 3 with slope minus 4, and g starting at 1 with slope 6. Multiply them out, writing u for x minus 2, and the coefficient of u is 14. Any functions with the same four numbers give the same answer.

32. Example 3.24: a product of binomials

Worked example

\[ j(x) = \left(x^2 + 2\right)\left(3x^3 - 5x\right) \]

Name the factors and differentiate each

Why: Power rule on each factor.

\[ f = x^2 + 2, \quad f' = 2x, \qquad g = 3x^3 - 5x, \quad g' = 9x^2 - 5 \]

Apply the product rule

Why: First derivative times second, plus second derivative times first.

\[ j'(x) = (2x)\left(3x^3 - 5x\right) + \left(9x^2 - 5\right)\left(x^2 + 2\right) \]

Expand each product

Why: Distribute.

\[ j'(x) = 6x^4 - 10x^2 + 9x^4 + 18x^2 - 5x^2 - 10 \]

Collect like terms

Why: Fourth powers, then squares.

\[ j'(x) = 15x^4 + 3x^2 - 10 \]

Check by multiplying out first

Why: Expand j, then use only the power rule.

\[ j(x) = 3x^5 + x^3 - 10x \;\Rightarrow\; j'(x) = 15x^4 + 3x^2 - 10 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 222 — Example 3.24

Name the two factors, differentiate each one, then assemble. Writing f, f prime, g and g prime on one line before you assemble keeps the substitution tidy and is the single best habit for product-rule problems.

The expanding is where marks are lost, so the slide separates it from the collecting. There are six terms after distributing; the x squared terms, minus ten, eighteen and minus five, combine to three x squared.

The book's check is the best kind: do the problem a completely different way. Multiplying the binomials first gives three x to the fifth plus x cubed minus ten x, and the power rule on that gives the same answer. When both routes agree, you can be confident in both.

33. Checkpoint 3.16: two x to the fifth times a quadratic

Worked example

\[ j(x) = 2x^5\left(4x^2 + x\right) \]

Name the factors and differentiate each

Why: Constant multiple and power rules.

\[ f = 2x^5, \quad f' = 10x^4, \qquad g = 4x^2 + x, \quad g' = 8x + 1 \]

Apply the product rule

Why: Two terms.

\[ j'(x) = 10x^4\left(4x^2 + x\right) + (8x + 1)\left(2x^5\right) \]

Expand

Why: Distribute each product.

\[ j'(x) = 40x^6 + 10x^5 + 16x^6 + 2x^5 \]

Collect

Why: Like powers.

\[ j'(x) = 56x^6 + 12x^5 \]

Check by expanding j first

Why: Multiply out, then use the power rule.

\[ j(x) = 8x^7 + 2x^6 \;\Rightarrow\; j'(x) = 56x^6 + 12x^5 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 223 — Checkpoint 3.16

Same method, and again it is worth writing the four ingredients before you assemble: the first factor, two x to the fifth, has derivative ten x to the fourth, and the second, four x squared plus x, has derivative eight x plus 1.

After distributing you have two x to the sixth terms, forty and sixteen, and two x to the fifth terms, ten and two. Collecting gives fifty-six x to the sixth plus twelve x to the fifth.

In a problem like this the product rule is not the fastest route: multiplying out first gives eight x to the seventh plus two x to the sixth, whose derivative you can write down immediately. The checkpoint asks for the product rule to give you practice, and the expansion makes the perfect check.

34. Trap: multiplying the derivatives

Trap

The trap

A product differentiated as if it were a sum:

\[ \frac{d}{dx}\left(x^2\cdot x^3\right) = 2x\cdot 3x^2 \]

\[ = 6x^3 ? \]

Wrong. Wrong coefficient and wrong degree.

The fix

A product changes in two ways at once, so its derivative has two terms.

\[ 2x\cdot x^3 + 3x^2\cdot x^2 = 5x^4 \]

Quick test: x squared times x cubed is x to the fifth, whose derivative is five x to the fourth. A degree check alone exposes the wrong line.

This is the most common error in the whole lesson, and it survives because the sum rule trained you to treat each piece separately. A product is different: both factors change at once, and each change is multiplied by the other factor.

The fastest way to catch it is a degree check. x squared times x cubed has degree five, so its derivative must have degree four. The wrong answer, six x cubed, has degree three. Whenever you differentiate a product, glance at the degree of your answer; one line of mental arithmetic exposes the wrong rule.

35. The quotient rule

Section

Part 4

36. Break the quotient guess

Counterexample

Discussion prompt

Someone guesses that the derivative of f over g is f prime over g prime. Find a quotient you can already differentiate another way that breaks the guess.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 223 — The Quotient Rule, introduction

Try to find the counterexample yourself before revealing. The strategy is the same one that broke the product guess: choose a quotient that simplifies to something you already know how to differentiate.

x cubed over x simplifies to x squared, whose derivative is two x. The guess, derivative of the top over derivative of the bottom, gives three x squared over 1. Those disagree for every x except 0 and two thirds, so the guess is wrong.

One counterexample is enough to kill a proposed rule. It also tells you what the true rule must do: it has to involve the top and bottom themselves, not only their derivatives, just as the product rule did.

37. Deriving the quotient rule from the product rule

Concept

The book omits the proof. Here is a short one, assuming the quotient is differentiable: write the quotient as q, so the numerator is a product.

\[ q = \frac{f}{g} \quad\Longrightarrow\quad f = q\,g \]

\[ f' = q'g + qg' \]

\[ q' = \frac{f' - qg'}{g} = \frac{f' - \frac{f}{g}g'}{g} \]

\[ q' = \frac{f'g - fg'}{g^2} \]

Quotient rule (Theorem 3.6) — The derivative of a quotient is the derivative of the top times the bottom, minus the derivative of the bottom times the top, all over the bottom squared.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 223 — Theorem 3.6

The book states the quotient rule without proof, saying it is similar to the product rule. Here is a quicker route that uses the product rule you just proved, assuming the quotient is differentiable.

Call the quotient q. Then the top, f, equals q times g, which is a product. Differentiate both sides with the product rule, and you get f prime equals q prime g plus q g prime. Now solve that equation for q prime: subtract q g prime and divide by g.

Finally replace q by f over g and clear the fraction inside the fraction by multiplying top and bottom by g. The result is the quotient rule. Seeing it come out of the product rule explains its shape: the same two pieces, a minus sign from moving one of them across, and a g squared from dividing twice.

38. Reading the quotient rule piece by piece

Notation

Annotate

On: \( \frac{d}{dx}\left(\frac{f}{g}\right) = \frac{f'(x)g(x) - g'(x)f(x)}{\left(g(x)\right)^2} \)

  • The derivative of the TOP comes first. Start with the top, as when reading the fraction.
  • Unlike the product rule this is a difference, so swapping the two terms flips the sign of the answer.
  • The derivative of the bottom times the top: the same pieces as the product rule.
  • The bottom squared. Leave it unexpanded; it is positive wherever g is not zero, which helps when reading the sign.

The rule has a minus sign, so order matters. The derivative of the top comes first. One way to remember it is to follow the fraction as you read it: top first, so the top's derivative leads.

The bottom squared in the denominator deserves a comment. Leave it unexpanded. It is never negative, so the sign of the whole derivative is the sign of the numerator, which makes it easy to check your answer against a graph.

The same four ingredients as the product rule appear: the top, the bottom, and their derivatives. Write all four down before you assemble, exactly as you did for products.

39. Example 3.25: five x squared over 4x + 3

Worked example

\[ k(x) = \frac{5x^2}{4x + 3} \]

Name top and bottom and differentiate each

Why: Power rule on each.

\[ f = 5x^2, \quad f' = 10x, \qquad g = 4x + 3, \quad g' = 4 \]

Substitute into the quotient rule

Why: Top derivative first.

\[ k'(x) = \frac{10x(4x + 3) - 4\left(5x^2\right)}{(4x + 3)^2} \]

Expand the numerator

Why: Distribute.

\[ k'(x) = \frac{40x^2 + 30x - 20x^2}{(4x + 3)^2} \]

Collect

Why: Leave the denominator squared.

\[ k'(x) = \frac{20x^2 + 30x}{(4x + 3)^2} \]

Check at x equal to 0

Why: k is zero at 0 and nonnegative just either side, so its graph has a horizontal tangent there; the formula gives slope zero.

\[ k'(0) = \frac{0 + 0}{9} = 0 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 223 — Example 3.25

Set up the four ingredients: top five x squared with derivative ten x, bottom four x plus 3 with derivative 4. Then substitute into the rule with the top's derivative first.

The simplification is short but needs care with the second product: four times five x squared is twenty x squared, and it is subtracted. The numerator collects to twenty x squared plus thirty x, and the denominator stays as a square.

The check uses the shape of the function. At x equal to 0, k is zero, and just either side of 0 it is positive, because a square over a positive number is positive. So the graph dips to touch zero at the origin and the tangent there must be horizontal. The formula agrees: the derivative at 0 is zero.

40. Checkpoint 3.17: a quotient of two lines

Worked example

\[ h(x) = \frac{3x + 1}{4x - 3} \]

Top, bottom and their derivatives

Why: Both are linear.

\[ f = 3x + 1, \quad f' = 3, \qquad g = 4x - 3, \quad g' = 4 \]

Quotient rule

Why: Top derivative first.

\[ h'(x) = \frac{3(4x - 3) - 4(3x + 1)}{(4x - 3)^2} \]

Expand the numerator

Why: The x terms cancel.

\[ h'(x) = \frac{12x - 9 - 12x - 4}{(4x - 3)^2} \]

Simplify

Why: A constant over a square.

\[ h'(x) = \frac{-13}{(4x - 3)^2} \]

Figure (svg): The graph of h of x equals 3x plus 1 over 4x minus 3: two decreasing branches either side of the vertical asymptote x equals three quarters, levelling toward y equals three quarters, with the tangent line at the point (2, 1.4) of slope minus 0.52.

Both branches fall from left to right, exactly as a derivative that is negative everywhere promises.

Check the sign and a value

Why: The derivative is negative wherever it exists, and the graph falls on both branches. At 2 the formula gives minus 0.52; a small step gives about the same.

\[ h'(2) = \frac{-13}{25} = -0.52, \qquad \frac{h(2.001) - h(2)}{0.001} \approx -0.5196 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 223 — Checkpoint 3.17

Both top and bottom are linear, so their derivatives are the constants 3 and 4. After substituting, something pleasant happens in the numerator: the twelve x terms cancel, leaving the constant minus 13.

A constant numerator over a square tells you a lot. The square is positive wherever the function is defined, so the derivative is negative everywhere. The graph must fall on both sides of the vertical asymptote, and the figure shows exactly that.

The numerical check at x equal to 2 gives a slope of about minus 0.5196 for a small step, matching the exact value minus 0.52. The dashed tangent line in the figure has that slope.

41. Find the error: the terms swapped

Error analysis

Annotate

On: \( \frac{d}{dx}\left(\frac{1}{x}\right) = \frac{g'f - f'g}{g^2} = \frac{(1)(1) - (0)(x)}{x^2} = \frac{1}{x^2} \)

  • The terms are in the wrong order. The rule is f′g − g′f: top derivative first.
  • With f = 1 and g = x, the correct numerator is (0)(x) − (1)(1) = −1.
  • Sign check: y = 1/x falls for every x > 0, so its slope must be negative. The answer here is positive, so it cannot be right. The correct derivative is −1/x².

Read the argument and try to spot the mistake before revealing the annotations. The arithmetic after the first equals sign is fine; the error is in the formula itself.

The numerator was written with the derivative of the bottom first. Because the rule has a minus sign, swapping the terms multiplies the whole answer by minus one. The correct numerator is zero times x minus one times one, which is minus one.

The last annotation gives you a check you can use on every quotient-rule problem. One over x decreases for every positive x, so its derivative must be negative there. A positive answer cannot be right. Comparing the sign of your derivative with the direction of the graph takes seconds and catches this error every time.

42. Theorem 3.7: negative exponents

Concept

The quotient rule extends the power rule to negative integers. Let k be a negative integer and write n for minus k, a positive integer.

\[ x^{k} = x^{-n} = \frac{1}{x^n}, \qquad f = 1,\; f' = 0, \qquad g = x^n,\; g' = nx^{n-1} \]

\[ \frac{d}{dx}\left(x^{-n}\right) = \frac{0\cdot x^n - 1\cdot nx^{n-1}}{\left(x^n\right)^2} = \frac{-nx^{n-1}}{x^{2n}} \]

\[ = -nx^{(n-1) - 2n} = -nx^{-n-1} \]

\[ \frac{d}{dx}\left(x^{k}\right) = kx^{k-1} \quad \text{for every negative integer } k \]

The same formula as before: the power rule now holds for every integer exponent.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 223-224 — Theorem 3.7 and its proof

The power rule was proved only for positive whole exponents, because it came from the binomial theorem. The quotient rule extends it to negative exponents with almost no effort.

Write x to a negative power as one over x to a positive power. The top is the constant 1, with derivative zero, so the first term of the quotient rule vanishes, leaving minus n x to the n minus 1, over x to the two n. Subtracting exponents gives minus n x to the minus n minus 1.

Now rename. With k equal to minus n, the coefficient minus n is just k, and the exponent minus n minus 1 is k minus 1. The formula is exactly the power rule. The rule you already knew now works for every integer exponent, positive, negative or zero.

43. Examples 3.26 and 3.27: negative powers

Worked example

\[ \text{(a) } \frac{d}{dx}\left(x^{-4}\right) \qquad \text{(b) } f(x) = \frac{6}{x^2} \]

Apply the extended power rule to (a)

Why: k is minus 4, so the new exponent is minus 5.

\[ \frac{d}{dx}\left(x^{-4}\right) = -4x^{-4-1} = -4x^{-5} \]

Rewrite (b) as a power

Why: Far easier than the quotient rule.

\[ f(x) = \frac{6}{x^2} = 6x^{-2} \]

Constant multiple rule

Why: Keep the 6 aside.

\[ f'(x) = 6\frac{d}{dx}\left(x^{-2}\right) \]

Extended power rule

Why: k is minus 2.

\[ f'(x) = 6\left(-2x^{-3}\right) \]

Simplify

Why: Multiply the constants.

\[ f'(x) = -12x^{-3} = -\frac{12}{x^3} \]

Check (b) with the quotient rule

Why: The longer route must agree; Checkpoint 3.18 works like (a), giving minus 7 x to the minus 8 for one over x to the seventh.

\[ \frac{0\cdot x^2 - 2x\cdot 6}{x^4} = \frac{-12x}{x^4} = -\frac{12}{x^3} \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 224 — Examples 3.26 and 3.27, Checkpoint 3.18

For part (a) the rule is immediate: bring down minus 4 and lower the exponent to minus 5. Take care with that lowering. Going down by one from minus 4 means minus 5, not minus 3; this is the most common slip with negative exponents.

Part (b) is the example the book uses to make a point about strategy. Six over x squared is a quotient, and the quotient rule would work, but rewriting it as six x to the minus 2 turns it into a one-line power-rule problem. Whenever the top is a constant and the bottom is a single power, rewrite.

The check does the problem the long way with the quotient rule, and the answers agree. Checkpoint 3.18 is the same as part (a): one over x to the seventh is x to the minus 7, with derivative minus seven x to the minus 8.

44. Does the power rule survive a fractional exponent?

Prediction

\[ \sqrt{x} = x^{1/2} \]

Predict first

Apply the power rule to x^(1/2) as if it were allowed. Does the result match the derivative of √x found from the definition in Section 3.2?

  • Yes: it gives 1/(2√x)
  • No: it gives (1/2)√x
  • No: the rule only makes sense for whole numbers
  • No: it gives 2√x

Correct: Yes: it gives 1/(2√x)

Why: Formally, (1/2)x^(1/2 − 1) = (1/2)x^(−1/2) = 1/(2√x), exactly the answer the conjugate calculation produced. The book proves the rule for rational exponents later; for now this agreement is strong evidence.

\[ \frac{d}{dx}\left(x^{1/2}\right) = \tfrac12 x^{-1/2} = \frac{1}{2\sqrt x} \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 217-218 — the remark after Example 3.18

Make your prediction before revealing. The power rule has only been proved for integer exponents so far, so it is a fair question whether it means anything for an exponent of one half.

Apply it formally: bring down one half and lower the exponent by one, to minus one half. x to the minus one half is one over the square root of x, so the result is one over two root x. That is exactly the derivative the book found from the definition with the conjugate trick.

Agreement on one example is evidence, not proof. The book extends the power rule to rational exponents later in the chapter. Until then you may use this result for the square root, because it was proved separately from the definition.

45. Function to derivative

Matching

Match the pairs

  • a. x⁷ + 10
  • b. 1/x³
  • c. √x
  • d. 4x² − 7x
  • e. x⁴ + 2/x
  • p. 7x⁶
  • q. −3/x⁴
  • r. 1/(2√x)
  • s. 8x − 7
  • t. 4x³ − 2/x²

Why: Rewrite first, then use the power rule: 1/x³ = x⁻³ gives −3x⁻⁴; √x = x^(1/2) gives (1/2)x^(−1/2); 2/x = 2x⁻¹ gives −2x⁻². The constant 10 contributes nothing. Items a, d and e are Exercises 106, 108 and 110.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 228 — Exercises 106, 108 and 110

Match each function to its derivative. For most of them the key move comes before any differentiating: rewrite roots and reciprocals as powers. One over x cubed is x to the minus 3; the square root is x to the one half; two over x is two x to the minus 1.

Then the power rule does everything. Watch the negative exponents in particular: the derivative of two x to the minus 1 is minus two x to the minus 2, so the plus sign in x to the fourth plus two over x becomes a minus sign in the derivative. And the constant 10 in the first function simply disappears.

46. Combining the rules

Section

Part 5

47. Undo the order of evaluation

Concept

Figure (svg): An expression tree for h of x equals 2x cubed k of x over 3x plus 2. The root is a division node. Its left child is a multiplication node whose children are 2x cubed and k of x; its right child is 3x plus 2.

Evaluating h works from the leaves up; differentiating works from the root down, so the first rule you use is the last operation you would compute.

To evaluate h at a number you would compute two x cubed times k, then three x plus two, then divide. The rules are applied in the reverse order: the division was last, so the quotient rule comes first, and the product rule is used inside it.

\[ h(x) = \frac{2x^3k(x)}{3x + 2} \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 224-225 — Combining Differentiation Rules

When several rules are needed, the question is which one to use first. The book's rule of thumb is to apply the rules in the reverse of the order in which you would evaluate the function. The tree in the figure makes that concrete.

To evaluate h at a number, you would start at the bottom of the tree: compute two x cubed and k of x, multiply them, compute three x plus two, and finally divide. Division is the last step, so it sits at the root of the tree. Differentiating starts at the root: the quotient rule first, and then, inside it, the product rule for the top.

Ask yourself, for any expression: if I had to evaluate this with a calculator, what would be the very last button I press? That operation tells you the first rule.

48. Example 3.28: a sum containing a product

Worked example

\[ k(x) = 3h(x) + x^2g(x) \]

Sum rule first

Why: The last operation is addition.

\[ k'(x) = \frac{d}{dx}\left(3h(x)\right) + \frac{d}{dx}\left(x^2g(x)\right) \]

Constant multiple rule on the first term

Why: The 3 stays.

\[ \frac{d}{dx}\left(3h(x)\right) = 3h'(x) \]

Product rule on the second term

Why: Factors x squared and g.

\[ \frac{d}{dx}\left(x^2g(x)\right) = 2x\,g(x) + x^2g'(x) \]

Assemble

Why: Add the pieces.

\[ k'(x) = 3h'(x) + 2xg(x) + x^2g'(x) \]

Check with h and g both equal to x

Why: Then k is three x plus x cubed, whose derivative is known directly.

\[ 3(1) + 2x(x) + x^2(1) = 3 + 3x^2 = \frac{d}{dx}\left(3x + x^3\right) \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 225 — Example 3.28

The last operation in k is the addition, so the sum rule comes first and splits the problem into two independent pieces. The first piece, three h, is a constant times a function, so the constant multiple rule applies. The second, x squared times g, is a product of two functions of x, so it needs the product rule.

Notice the difference between the two pieces. The 3 is a constant and just rides along. The x squared is a function of x and changes, so it gets differentiated in one of the product rule's two terms.

The check replaces the unknown functions by simple ones you can handle directly. With h and g both equal to x, k becomes three x plus x cubed, whose derivative is three plus three x squared, and the general formula gives exactly that.

49. Checkpoint 3.19: three f minus two g

Prediction

\[ \frac{d}{dx}\left(3f(x) - 2g(x)\right) = \;? \]

Predict first

Which rules apply, and what is the derivative?

  • 3f′(x) − 2g′(x)
  • 3f′(x) · (−2g′(x))
  • f′(x) − g′(x)
  • 3f′(x) − 2g′(x) + 1

Correct: 3f′(x) − 2g′(x)

Why: The last operation is subtraction, so the difference rule splits it; then the constant multiple rule keeps the 3 and the 2. No product rule is needed, because 3 and 2 are constants, not functions of x.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 225 — Checkpoint 3.19

Choose an answer before revealing. The only trap here is reaching for the product rule because you see a number multiplied by a function.

The product rule is for a product of two functions that both change with x. A number like 3 does not change, so the constant multiple rule applies instead: it just stays in front. The difference rule splits the expression, and the answer is three f prime minus two g prime.

If you did use the product rule on three f, you would get zero times f plus three f prime, which is the same answer. So the product rule is not wrong here, just unnecessary. The constant multiple rule is its shortcut.

50. Example 3.29: a product of three functions

Worked example

Express the derivative of k in terms of f, g, h and their derivatives.

\[ k(x) = f(x)g(x)h(x) \]

Group two of the factors

Why: Treat the product fg as one function.

\[ k = (fg)\cdot h \]

Product rule on the outer product

Why: Differentiate fg, then h.

\[ k' = (fg)'h + h'(fg) \]

Product rule on fg

Why: Replace the derivative of fg.

\[ k' = (f'g + g'f)h + h'fg \]

Distribute

Why: One term for each factor that is differentiated.

\[ k' = f'gh + fg'h + fgh' \]

Check with f, g, h equal to x, x + 1, x + 2 at x equal to 1

Why: The product is x cubed plus three x squared plus two x, with derivative 11 at 1; the formula also gives 11.

\[ 1\cdot 2\cdot 3 + 1\cdot 1\cdot 3 + 1\cdot 2\cdot 1 = 11 = 3 + 6 + 2 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 225 — Example 3.29

The product rule is for two factors, so the trick is to group. Treat f times g as a single function and multiply it by h. The product rule on that gives the derivative of f g times h, plus h prime times f g.

The derivative of f g needs the product rule again. Substitute it and distribute, and a pattern appears: three terms, and in each one exactly one of the three factors is differentiated. With four factors you would get four terms, and so on.

The check uses three simple linear factors, x, x plus 1 and x plus 2, at x equal to 1. The formula gives six plus three plus two, which is eleven. Multiplying out gives x cubed plus three x squared plus two x, whose derivative at 1 is three plus six plus two, also eleven.

51. Example 3.30: a product inside a quotient

Worked example

\[ h(x) = \frac{2x^3k(x)}{3x + 2} \]

Quotient rule first

Why: Division is the last operation.

\[ h'(x) = \frac{\frac{d}{dx}\left(2x^3k(x)\right)(3x+2) - \frac{d}{dx}(3x+2)\left(2x^3k(x)\right)}{(3x+2)^2} \]

Product rule for the top

Why: Factors two x cubed and k.

\[ \frac{d}{dx}\left(2x^3k(x)\right) = 6x^2k(x) + 2x^3k'(x) \]

Derivative of the bottom

Why: A line of slope 3.

\[ \frac{d}{dx}(3x + 2) = 3 \]

Substitute

Why: Numerator only; the denominator stays.

\[ \text{numerator} = \left(6x^2k + 2x^3k'\right)(3x+2) - 3\left(2x^3k\right) \]

Expand

Why: Five terms.

\[ = 18x^3k + 12x^2k + 6x^4k' + 4x^3k' - 6x^3k \]

Collect and factor

Why: The k terms and the k prime terms separately.

\[ h'(x) = \frac{12k(x)\left(x^3 + x^2\right) + 2k'(x)\left(3x^4 + 2x^3\right)}{(3x+2)^2} \]

Check with k equal to 1

Why: Then k prime is zero and the formula must match the quotient rule on two x cubed over three x plus two.

\[ \frac{\left(6x^2\right)(3x+2) - 3\left(2x^3\right)}{(3x+2)^2} = \frac{12x^3 + 12x^2}{(3x+2)^2} \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 225 — Example 3.30

This example is the tree from a few slides ago in action. The quotient rule goes first, and at first you leave the derivative of the top as an instruction, d over d x of two x cubed k. Then you compute that piece separately with the product rule, and the bottom's derivative is just 3.

Substituting and expanding gives five terms. Three of them contain k and two contain k prime. The k terms, eighteen x cubed k, twelve x squared k and minus six x cubed k, combine to twelve x cubed k plus twelve x squared k, and factoring gives the book's final form.

The check sets k equal to 1, which makes k prime zero. Then h is just two x cubed over three x plus two, which you can differentiate directly with the quotient rule, and it matches the general formula with k equal to 1.

52. Order the attack on Example 3.30

Ranking

Put in order

Put the steps for differentiating h(x) = 2x³k(x)/(3x + 2) in order.

  1. Identify the last operation: a division
  2. Write the quotient rule with the top's derivative still unevaluated
  3. Use the product rule to differentiate the top, 2x³k(x)
  4. Differentiate the bottom: 3
  5. Expand the numerator and collect the k and k′ terms

Why: Outside in: the outermost operation decides the first rule, the pieces it needs are differentiated next, and simplifying comes last. Differentiating the bottom could swap with the product rule step, but both must come before expanding.

Put the steps in order before checking. The first step is not a calculation at all: it is recognising the structure. Everything else follows from knowing that the outermost operation is division.

Writing the quotient rule with the top's derivative still unevaluated is a deliberate move. It keeps the structure visible on the page, so you do not lose track of which piece goes where. Then you fill in each piece, and only at the very end do you expand and simplify.

53. Which rule comes first?

Sorting

Sort into buckets

Sort each function by the FIRST rule you would use on it (after any rewriting).

Product rule
(x + 2)(2x² − 3); 3x(18x⁴ + 13/(x + 1))
Quotient rule
(x² + 4)/(x² − 4); (x + 9)/(x² − 7x + 1)
Sum rule, termwise
8x⁴ + 9x² − 1; 4x² − 7x
prod
The last operation is multiplication of two non-constant factors. (In d the second factor itself contains a quotient, handled inside.)
quot
The last operation is division by a non-constant function of x, so the quotient rule is needed.
sum
A polynomial written as a sum of terms: split it and use the power and constant multiple rules.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 228 — Exercises 108, 109, 111, 112, 116, 117

For each function, imagine evaluating it at a number and ask what the last operation would be. That operation decides the first rule. These functions are all from the section's exercise set.

Two of them are sums of terms, so they are differentiated termwise. Two are quotients of non-constant functions, so they need the quotient rule. The remaining two are products; one of them, three x times a bracket that contains a fraction, needs the quotient rule as well, but only inside the product rule, as a second step.

Sorting like this before you calculate is the habit that makes combined problems manageable. It turns one big problem into a sequence of small ones.

54. Before reaching for the quotient rule

Step zero

\[ f(x) = \frac{\left(x^2 + 1\right)(x - 2)}{x} \]

Discussion prompt

This looks like a product inside a quotient. Before you use either rule, what could you do that makes the derivative a one-line power-rule calculation? Do it.

Write your plan before revealing. The function looks like it needs both the product rule and the quotient rule, and that route works, but it is long and easy to get wrong.

The denominator is a single term, x. That means you can divide every term of the numerator by it. Multiply out the top to get x cubed minus two x squared plus x minus 2, then divide each term by x. You are left with a polynomial plus a single negative power, and the power rule finishes it in one line.

Simplifying before differentiating is almost always worth a moment's thought. The rules will always get you there, but the simplest correct route is the one with the fewest places to make an error.

55. Flat tangents, parallel tangents, velocity

Section

Part 6

56. What a horizontal tangent means

Intuition

A horizontal line has slope zero. So the graph of f has a horizontal tangent exactly where the derivative is zero, and finding those points is an equation-solving problem.

\[ \text{horizontal tangent at } x = a \iff f'(a) = 0 \]

The same idea finds a tangent parallel to any given line: set the derivative equal to that line's slope.

\[ \text{tangent parallel to } y = mx + b \iff f'(a) = m \]

Up to now you have used the derivative to find slopes at given points. This part turns the question around: given a slope, find the points that have it.

The most important case is slope zero. A horizontal tangent line appears at the top of a hump or the bottom of a dip in a graph, and those are exactly the places where the derivative equals zero. So finding horizontal tangents means solving the equation f prime of x equals zero.

The same idea works for any slope. Parallel lines have equal slopes, so a tangent line parallel to a given line appears wherever the derivative equals that line's slope. Chapter 4 builds a whole theory of maxima and minima on this idea.

57. Example 3.31: where the tangent is horizontal

Worked example

\[ f(x) = x^3 - 7x^2 + 8x + 1 \]

Differentiate

Why: Termwise.

\[ f'(x) = 3x^2 - 14x + 8 \]

Set the derivative to zero

Why: Horizontal means slope zero.

\[ 3x^2 - 14x + 8 = 0 \]

Factor

Why: Look for factors of 3 times 8 that add to minus 14: minus 2 and minus 12.

\[ (3x - 2)(x - 4) = 0 \]

Solve

Why: Either factor may vanish.

\[ x = \frac23 \quad\text{or}\quad x = 4 \]

Find the points

Why: Evaluate f; the book gives only the x values.

\[ f\left(\tfrac23\right) = \frac{95}{27} \approx 3.52, \qquad f(4) = -15 \]

Figure (svg): The cubic f of x equals x cubed minus 7 x squared plus 8x plus 1 for x from minus 1.2 to 6.3, with short horizontal tangent segments at its local maximum near (0.67, 3.52) and its local minimum at (4, minus 15).

The tangent is flat exactly at the top of the hump and the bottom of the dip, the two places where the derivative is zero.

Check the factoring

Why: Multiply back, and confirm the derivative vanishes at two thirds.

\[ (3x-2)(x-4) = 3x^2 - 14x + 8, \qquad f'\left(\tfrac23\right) = \tfrac43 - \tfrac{28}{3} + 8 = 0 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 226 — Example 3.31 and Figure 3.19

Differentiate, set the derivative equal to zero, and solve. The derivative is a quadratic, so factoring is the natural tool. You need two numbers that multiply to three times eight, which is 24, and add to minus 14; minus 2 and minus 12 work, and they lead to the factors three x minus 2 and x minus 4.

The book stops at the x values, two thirds and 4. The slide also finds the heights, because a point on a graph needs both coordinates: about 3.52 at two thirds, and minus 15 at 4.

Look at the figure. The flat segments sit exactly at the top of the hump and the bottom of the dip. The check multiplies the factors back out to recover the derivative, and confirms the derivative is zero at two thirds.

58. Checkpoint 3.20: a tangent parallel to y = 2x + 3

Worked example

\[ f(x) = 4x^2 - 3x + 2 \]

Differentiate

Why: Termwise.

\[ f'(x) = 8x - 3 \]

Match the slope of the line

Why: Parallel lines have equal slopes, and the line has slope 2.

\[ 8x - 3 = 2 \]

Solve

Why: Add 3, divide by 8.

\[ x = \frac58 \]

Find the point and the tangent line

Why: Evaluate f at five eighths.

\[ f\left(\tfrac58\right) = \frac{27}{16}, \qquad y = 2x + \frac{7}{16} \]

Figure (svg): The parabola f of x equals 4 x squared minus 3x plus 2, the line y equals 2x plus 3, and the tangent line to the parabola at x equals five eighths, which runs parallel to that line.

Only one point on the parabola has slope 2; its tangent line is the dashed line, parallel to the given one.

Check

Why: The derivative at five eighths is 2, and the tangent line's intercept, seven sixteenths, differs from 3, so the lines are parallel and distinct.

\[ f'\left(\tfrac58\right) = 8\cdot\tfrac58 - 3 = 2 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 226 — Checkpoint 3.20

Parallel lines have the same slope, and the given line has slope 2. So you are looking for the points where the derivative equals 2, not zero.

The derivative is eight x minus 3, and setting it equal to 2 gives x equal to five eighths. There is only one such point, because the derivative of a parabola is a line, and a non-horizontal line takes each value exactly once.

The figure shows the tangent at that point running parallel to the given line, and the slide also finds its equation: its intercept is seven sixteenths, not 3, so the two lines are parallel and not the same line.

59. Example 3.32: an initial velocity

Worked example

The position of an object at time t is given below. Find its initial velocity.

\[ s(t) = \frac{t}{t^2 + 1} \]

Velocity is the derivative of position

Why: Initial means t equal to 0.

\[ v(0) = s'(0) \]

Quotient rule

Why: Top t, bottom t squared plus 1.

\[ s'(t) = \frac{1\cdot\left(t^2 + 1\right) - 2t\cdot t}{\left(t^2 + 1\right)^2} \]

Simplify

Why: Collect the t squared terms.

\[ s'(t) = \frac{1 - t^2}{\left(t^2 + 1\right)^2} \]

Evaluate at t equal to 0

Why: Numerator 1, denominator 1.

\[ v(0) = \frac{1 - 0}{1} = 1 \]

Figure (svg): The position curve s of t equals t over t squared plus 1 for t from minus 4 to 4, rising through the origin with the tangent line s equals t drawn there, then peaking at t equals 1.

At the start the object moves at unit speed; it slows, stops at t equal to 1, and drifts back toward the origin.

Check with a small time step

Why: Over the first thousandth of a unit of time the object moves almost exactly one thousandth.

\[ \frac{s(0.001) - s(0)}{0.001} = 0.999999 \approx 1 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 226 — Example 3.32

In physics terms, velocity is the rate of change of position, so it is the derivative of the position function. Initial velocity means the velocity at time zero.

The position is a quotient, so the quotient rule applies: top t with derivative 1, bottom t squared plus 1 with derivative two t. The numerator simplifies to one minus t squared. At time zero the numerator and the denominator are both 1, so the initial velocity is 1.

The figure shows more than the question asks. The tangent at the origin has slope 1. The derivative is zero at t equal to 1, so the object stops there, at the peak of the curve, and then moves back toward the origin. The small-step check gives almost exactly 1.

60. Exercise 144: when the car stops

Worked example

A car in traffic has travelled s metres after t seconds. Find when its velocity is zero, and its acceleration then.

\[ s(t) = t^3 - 6t^2 + 9t \]

Differentiate for velocity

Why: Power rule termwise.

\[ v(t) = s'(t) = 3t^2 - 12t + 9 \]

Set velocity to zero and factor

Why: Take out the 3 first.

\[ 3\left(t^2 - 4t + 3\right) = 3(t - 1)(t - 3) = 0 \]

Solve

Why: Two instants.

\[ t = 1 \text{ s} \quad\text{or}\quad t = 3 \text{ s} \]

Differentiate again for acceleration

Why: The same rules, applied to the velocity.

\[ a(t) = v'(t) = 6t - 12 \]

Evaluate at each instant

Why: In metres per second squared.

\[ a(1) = -6, \qquad a(3) = 6 \]

Figure (svg): The position s of t equals t cubed minus 6 t squared plus 9t for t from 0 to 4.3 seconds: it rises to 4 metres at t equals 1, falls back to 0 at t equals 3, then rises again; flat tangent segments mark the two instants where the velocity is zero.

Where the car stops going forward the curve is bending down (negative acceleration); where it stops going backward it is bending up.

Check the sign change

Why: Between the two stops the velocity should be negative, and it is.

\[ v(2) = 12 - 24 + 9 = -3 < 0 \]

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 229 — Exercise 144

This exercise puts the rules to work twice. The velocity is the derivative of the position, and the acceleration is the derivative of the velocity, which is the second derivative of the position. The same power rule handles both.

Setting the velocity to zero gives a quadratic that factors after taking out 3: the car stops at 1 second and at 3 seconds. At 1 second the acceleration is minus 6 metres per second squared, so the car is braking to a stop; at 3 seconds it is plus 6.

The figure shows the story: the car moves forward, stops at 1 second, rolls back until 3 seconds, and moves forward again. The check confirms the velocity is negative between the two stops, which is the rolling back.

61. Formula One grandstands

Real world

Figure (svg): The track y equals x cubed plus 3 x squared plus x for x from minus 3.2 to 0.6, the grandstand corner at (minus 1.9, 2.8), the dashed line y equals 2.8, and two tangent lines: from (minus 2, 2) with slope 1, crossing y equals 2.8 at x equals minus 1.2, and from (minus 2.5, 0.625) with slope 4.75, crossing it at about minus 2.04.

A car leaving at the planned point flies past to the right of the grandstand; one that leaves earlier, on the steeper tangent, reaches the grandstand's line to its left.

Discussion prompt

A section of track follows y = x³ + 3x² + x, and the grandstand's front corner is at (−1.9, 2.8). A car that loses control flies off along the tangent line. (1) Where near the turn is the slope 1? (2) Where does that tangent cross y = 2.8, and is that clear of the grandstand? (3) Repeat for a car that loses control at (−2.5, 0.625).

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, pp. 226-228 — Student Project: Formula One Grandstands

This is the book's student project, and it uses nothing but the derivative of a cubic and the equation of a tangent line. Work through the three questions before revealing.

Setting the derivative equal to 1 gives two solutions, zero and minus 2; minus 2 is the one near the turn, at the point minus 2, 2. The tangent there is y equals x plus 4, which reaches the grandstand's height at x equal to minus 1.2. That is to the right of the grandstand corner at minus 1.9, so the flying car misses the stand.

Now suppose the car loses control earlier, at minus 2.5. The slope there is 4.75, much steeper, and that tangent reaches the grandstand's height at about minus 2.04, to the left of the corner. The spectators are in its path. The design should change, and the whole argument took three derivatives and two lines.

62. Putting it together

Section

Part 7

63. Pattern: differentiating a rational function

Pattern

Figure (svg): A flow diagram: first rewrite roots and reciprocals as powers and simplify; then find the last operation. A sum or constant multiple is split termwise; a product goes to the product rule; a quotient to the quotient rule; each piece is sent back through the same diagram until only powers and constants remain.

Rewrite first, then peel the expression from the outside in.
  1. Rewrite first. Roots and reciprocals become powers; a single-term denominator splits the fraction.
  2. Find the last operation you would perform to evaluate the function.
  3. Sum, difference, constant multiple: split and differentiate each piece.
  4. Product: f′g + fg′. Quotient: (f′g − g′f)/g², top derivative first.
  5. Send each piece round again until only powers and constants remain; simplify last.
  6. Check: a degree count, a sign against the graph, or a special case you can do another way.

This is the procedure the whole lesson has been building. Its first step is not a rule: rewrite. Roots and reciprocals become powers, and a fraction with a single-term denominator splits into separate terms. Many problems end right there.

Then look for the last operation, pick the matching rule, and send each piece back through the same process. The diagram's return arrow is the key idea: the rules are applied recursively, from the outside in, until only powers and constants are left.

The last step is a check. You have several: the degree should drop by one for a polynomial, the sign should match whether the graph rises or falls, and a special case you can do another way should agree.

64. The rules side by side

Comparison

Comparison matrix

rulederivative ofequals
constantc0
powerxⁿ (any integer n)nx^(n−1)
constant multiplek·f(x)k·f′(x)
sumf(x) + g(x)f′(x) + g′(x)
productf(x)·g(x)f′g + g′f
quotientf(x)/g(x)(f′g − g′f)/g²

Fill in the blanks from memory before checking. Seeing the rules side by side shows their pattern: the constant, power, multiple and sum rules each produce one piece per piece, while products and quotients mix the functions with their derivatives.

The power rule entry now says any integer n, because Theorem 3.7 extended it. The quotient rule entry is the one to write most carefully; say the order aloud as you write it, derivative of the top times the bottom, first.

65. Check: a negative power

Check

Check your understanding

What is the derivative of f(x) = 4/x³?

  • A. −12/x⁴ (correct)
  • B. 4/(3x²)
  • C. −12/x²
  • D. 12/x⁴

Answer: A

Why: Rewrite as 4x⁻³. The power rule with k = −3 gives 4(−3)x⁻⁴ = −12x⁻⁴ = −12/x⁴. The exponent goes DOWN by one, from −3 to −4.

Why B tempts people
This differentiates the denominator as if it were the whole function; rewrite as a power first.
Why C tempts people
The exponent moved the wrong way: from −3 the new exponent is −4, not −2.
Why D tempts people
The sign is lost: bringing down −3 makes the coefficient negative.

Rewrite first, so the problem becomes four times x to the minus 3. Then bring down minus 3 and lower the exponent by one, to minus 4.

Two of the wrong answers come from the two classic slips with negative exponents: moving the exponent up instead of down, and dropping the minus sign that comes down in front. A quick sign check helps: four over x cubed falls for positive x, so its derivative must be negative there.

66. Check: the quotient rule's order

Check

Check your understanding

What is the derivative of (x + 1)/(x − 1)?

  • A. 2/(x − 1)²
  • B. −2/(x − 1)² (correct)
  • C. 1
  • D. 2x/(x − 1)²

Answer: B

Why: Top derivative first: (1·(x − 1) − 1·(x + 1))/(x − 1)² = (x − 1 − x − 1)/(x − 1)² = −2/(x − 1)².

Why A tempts people
The terms of the numerator were swapped, which flips the sign.
Why C tempts people
That is the derivative of the top divided by the derivative of the bottom, which is not the quotient rule.
Why D tempts people
The numerator was not expanded correctly; the x terms cancel, leaving a constant.

Write down the four ingredients: top x plus 1, bottom x minus 1, and both derivatives equal to 1. Then substitute with the top's derivative first.

The numerator is x minus 1, minus x plus 1, which is minus 2. The swapped order gives plus 2, and the sign check exposes it: the graph of this function falls on each side of its asymptote, so the derivative must be negative.

67. Check: rules from a table

Check

Check your understanding

Given f(2) = 5, f′(2) = 7, g(2) = 3, g′(2) = 1, find h′(2) for h(x) = f(x)/g(x) (Exercise 127).

  • A. 7
  • B. −16/9
  • C. 16/9 (correct)
  • D. 26/9

Answer: C

Why: h′(2) = (f′(2)g(2) − g′(2)f(2))/g(2)² = (7·3 − 1·5)/3² = (21 − 5)/9 = 16/9.

Why A tempts people
This is f′(2)/g′(2), the quotient of the derivatives.
Why B tempts people
The numerator's terms are in the wrong order.
Why D tempts people
The numerator used a plus sign, as in the product rule.

OpenStax Calculus Volume 1, §3.3 Differentiation Rules §3.3, p. 228 — Exercise 127

This is Exercise 127 from the section, and it works like Example 3.23: you do not need formulas for f and g, only their values and derivatives at 2.

Take the numbers from the table: seven times three, minus one times five, all over three squared, which is sixteen ninths. Each wrong answer is one of the typical slips: dividing the derivatives, swapping the order, or using the product rule's plus sign.

68. Explain the two terms

Explain it to yourself

Discussion prompt

Using the growing rectangle, explain in two or three sentences why the product rule has exactly two terms, and why there is no third term for the corner.

Write your explanation before revealing, and use the picture in it. Being able to explain why the rule has two terms means you will never be tempted to multiply the derivatives again.

A good answer names the two strips, says which factor changes in each, and then deals with the corner: after dividing by h it is still a difference quotient times a change in g, and the change in g goes to zero. If you can also say that the continuity of g is what makes that change go to zero, you have reconstructed the proof.

69. Exit ticket

Exit ticket

\[ f(x) = \frac{x}{x^2 + 4} \]

Discussion prompt

Find every point where the graph of f has a horizontal tangent. Name the rule you used and give both coordinates.

This combines the quotient rule with the horizontal-tangent idea. Differentiate, then set the derivative to zero.

A fraction is zero exactly when its numerator is zero and its denominator is not. Here the denominator, x squared plus 4, all squared, is never zero, so the only condition is four minus x squared equal to zero. That gives x equal to 2 or minus 2.

Remember to give both coordinates: f of 2 is one quarter and f of minus 2 is minus one quarter. These are the highest and lowest points of the graph, which you will be able to prove in Chapter 4.

70. Recap

Recap

ruleformulawatch out for
constant(c)′ = 0a constant term vanishes, a constant factor stays
power(xⁿ)′ = nxⁿ⁻¹, any integer nneeds a variable base: not for 3ˣ
sum, multiple(f ± g)′ = f′ ± g′, (kf)′ = kf′differentiate termwise
product(fg)′ = f′g + g′ftwo terms, never f′g′
quotient(f/g)′ = (f′g − g′f)/g²top derivative first; check the sign

\[ \text{horizontal tangent} \iff f'(x) = 0 \]

Next, Section 3.4 reads derivatives as rates of change, and Section 3.5 adds the trigonometric functions to the list of building blocks.

Stewart, Calculus: Early Transcendentals 8e, §3.1 Derivatives of Polynomials and Exponential Functions §3.1, pp. 172-182 — the power, sum and constant multiple rules in Stewart

Stewart, Calculus: Early Transcendentals 8e, §3.2 The Product and Quotient Rules §3.2, pp. 183-189 — the product and quotient rules in Stewart

Five rules, each proved once from the definition, now replace the definition for every polynomial and rational function. The constant and power rules handle the building blocks; the sum, multiple, product and quotient rules handle the ways blocks are combined.

The right-hand column is the list of mistakes to guard against: the power rule on a variable exponent, multiplying derivatives, swapping the quotient's terms. Each one has a quick check: a variable-base test, a degree count, a sign check against the graph.

The same rules gave you tangent lines, horizontal tangents and velocities. The next sections add more building blocks, the trigonometric, exponential and logarithmic functions, and one more way of combining functions, composition, handled by the chain rule.

Sources

  1. OpenStax Calculus Volume 1, §3.3 Differentiation Rules — Strang & Herman, OpenStax / Rice University, CC BY-NC-SA 4.0, pp. 216-229
  2. Stewart, Calculus: Early Transcendentals 8e, §3.1 Derivatives of Polynomials and Exponential Functions — James Stewart, Cengage Learning, 2016, pp. 172-182
  3. Stewart, Calculus: Early Transcendentals 8e, §3.2 The Product and Quotient Rules — James Stewart, Cengage Learning, 2016, pp. 183-189

Want this taught 1-on-1? Alexander tutors Calculus I — $55/session, free consultation.

Book on Wyzant · Text (657) 465-8108