5.3 The Fundamental Theorem of Calculus

The Mean Value Theorem for Integrals, the area function and Part 1 proving its derivative is the integrand, Part 2 evaluating any definite integral from an antiderivative, the chain rule for variable limits, and the net change reading.

Subject: Calculus I · 70 slides · symbolic lesson

Open the interactive version of this deck

What this lesson covers

The lesson, slide by slide

1. The Fundamental Theorem of Calculus

Title

Calculus I · Section 5.3

Why an antiderivative computes an area

2. What this lesson gives you

Objectives

Sections 5.1 and 5.2 built the definite integral as a limit of Riemann sums. It is the right definition and a slow way to compute. This lesson connects the integral to the derivative, and the connection turns a page of limits into two substitutions.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 472-473 — learning objectives 5.3.1 to 5.3.6

So far you have two ways to find a definite integral: recognise a shape whose area you know, or take a limit of Riemann sums. The first works only for triangles, rectangles and circles. The second always works in principle and almost never in practice, because it needs a closed formula for a sum.

This lesson is the turning point of the whole course. It shows that the integral and the derivative, which you met as two unrelated ideas, are inverse operations. Once you see that, every derivative rule you learned in Chapter 3 becomes an integration rule for free.

The theorem has two parts, and you will use them for different jobs. Part 1 differentiates an integral whose upper limit is a variable. Part 2 evaluates a definite integral by finding an antiderivative. Before either, you need a small tool about averages that makes the proofs work.

3. Before anything new: two tools you already own

Warm-up

Discussion prompt

Write down, from memory, (1) the general antiderivative of x to the power n, and (2) what the definite integral from a to b of f(x) dx means as a limit.

Write both answers before you reveal them. If the power rule came back quickly but the Riemann sum definition did not, look at it again now, because the proofs in this lesson lean on it.

Notice how different the two statements feel. The first is about undoing a derivative: which function has x to the n as its slope? The second is about adding up thin rectangles: how much area sits under a curve? Nothing in either definition hints that they have anything to do with each other.

That surprise is the content of this lesson. By the end you will be able to compute the second thing using only the first, and you will see exactly why that is allowed.

4. The Mean Value Theorem for Integrals

Section

Part 1

5. The problem: Riemann sums are slow

Concept

Figure (svg): The parabola y equals x squared from 0 to 2 with eight right-endpoint rectangles drawn under and above it; their total area 3.1875 overshoots the true area 8 over 3.

Eight rectangles still miss by more than half a unit. A limit of such sums gives the exact answer, but only after a page of algebra.

With n right-endpoint rectangles on the interval from 0 to 2, the heights are the squares of 2i over n and each width is 2 over n.

\[ R_n = \sum_{i=1}^{n}\left(\frac{2i}{n}\right)^2 \frac{2}{n} = \frac{8}{n^3}\sum_{i=1}^{n} i^2 \]

\[ R_n = \frac{8}{n^3}\cdot\frac{n(n+1)(2n+1)}{6} \;\to\; \frac{16}{6} = \frac83 \]

It works, but only because a formula for the sum of squares exists. For most functions no such formula does. The rest of the lesson replaces this whole page with one subtraction.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 472 — introduction

Look at the figure. Eight rectangles, each taking its height from the right end of its slice, overshoot the true area by more than half a unit. To get the exact answer you have to let the number of rectangles grow without bound.

The algebra on the slide is the honest way to do that. You factor out the constants, recognise the sum of the first n squares, use its formula, and let n run off to infinity. The answer, eight thirds, is correct, and it took four lines and a formula you may not remember.

Now imagine doing this for the square root of x, or for sine. There is no convenient formula for those sums. That is the problem this section solves: by the end of Part 4 of this lesson, the same integral will take one line.

6. The average value of a function

Concept

Section 5.2 defined the average of a continuous function on an interval: the integral divided by the width.

\[ f_{\text{ave}} = \frac{1}{b-a}\int_a^b f(x)\,dx \]

\[ f_{\text{ave}}\,(b-a) = \int_a^b f(x)\,dx \]

Read the second line as geometry: a rectangle of height equal to the average, on the same base, has exactly the area under the curve. The average is the height that levels the region flat.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 473 — the average value formula

You met this formula at the end of Section 5.2. It generalises the everyday average: add up the values and divide by how many there are. For a function the adding up is the integral, and the how many is the width of the interval.

The second line is the one to picture. Multiply both sides by the width and the equation says: a rectangle whose height is the average, on the same base as the region, has exactly the same area as the region under the curve.

Think of the region as water in a tank with a wavy bottom turned upside down. Let it settle and it lies flat at one level. That level is the average value, and the next theorem says the curve must actually reach that level somewhere.

7. Theorem 5.3: a continuous function hits its average

Concept

Mean Value Theorem for Integrals (Theorem 5.3) — If f is continuous on the closed interval from a to b, then there is at least one point c in that interval where the value of f equals its average value.

\[ f(c) = \frac{1}{b-a}\int_a^b f(x)\,dx, \qquad c \in [a,b] \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 473 — Theorem 5.3, equation 5.15

The Mean Value Theorem for Integrals is a guarantee, not a method. It tells you a point c exists where the function equals its own average, but it does not tell you where. To find c, you compute the average and then solve an equation, as the examples will show.

The word continuous is essential, and so is the closed interval. You will see shortly that a function with a single jump can skip right over its average and never take that value.

Why name it after the Mean Value Theorem from Chapter 4? Because it is the same statement in disguise. If F is an antiderivative of f, the average of f is the slope of the secant line of F, and the ordinary Mean Value Theorem finds a point where F prime, which is f, equals that slope.

8. The levelling rectangle

Picture it

Figure (svg): A wavy continuous curve over the interval from 0 to 4, its region shaded, and a rectangle of the same base and equal area whose height is the average value, about 2. The curve crosses the top of the rectangle at three points c1, c2, c3.

The rectangle and the region have the same area. Where the curve crosses the top of the rectangle, the function equals its own average; here that happens three times.

Pour the region into a rectangle on the same base and it settles at the average height. The curve starts below that level and ends above it, so, being unbroken, it has to cross it somewhere.

Look at the yellow rectangle. It has the same base as the blue region and the same area, so its height is the average value, about 2. Wherever the curve crosses the top edge of the rectangle, the function equals its average. Here there are three such points.

The theorem only promises at least one. This curve happens to wiggle through the level three times, and each of the three crossing points is a legitimate answer for c.

Here is the intuition for why a crossing must exist. The rectangle's area equals the region's area, so the curve cannot sit entirely above the rectangle's top, or the region would have more area. It cannot sit entirely below either. So it is sometimes above and sometimes below, and an unbroken curve has to cross the level in between.

9. Reading Theorem 5.3 piece by piece

Notation

Annotate

On: \( \int_a^b f(x)\,dx = f(c)\,(b-a) \quad\text{for some } c \in [a,b] \)

  • The area under the curve (a signed area if f dips below the axis).
  • One actual value of the function, the height of the curve at the special point c.
  • The width of the interval. Height times width: the area of the levelling rectangle.
  • At least one, possibly several. The theorem promises existence, not a formula, and not uniqueness.

Step through the annotations and connect each piece to the picture you just saw. The integral is the area of the region; f of c times the width is the area of the levelling rectangle. The theorem says these two areas are equal for a suitable choice of c.

Pay attention to the last annotation. Some c means at least one, found by solving an equation. For a line there is exactly one; for a wavy curve there may be several; the theorem never tells you which.

If the function dips below the axis, the integral counts that part as negative, and the average can be negative too. The theorem still holds: a continuous function still takes its average value somewhere.

10. Why it is true: squeeze, then the Intermediate Value Theorem

Concept

A continuous function on a closed interval has a minimum m and a maximum M (the Extreme Value Theorem). Every height lies between them, so the area does too.

\[ m \le f(x) \le M \;\Longrightarrow\; m(b-a) \le \int_a^b f(x)\,dx \le M(b-a) \]

\[ m \le \frac{1}{b-a}\int_a^b f(x)\,dx \le M \]

The average is a number between the smallest and largest value of f. A continuous function takes every value between m and M, so it takes this one, at some point c.

\[ f(c) = f_{\text{ave}} \quad \text{(Intermediate Value Theorem)} \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 473 — proof of Theorem 5.3

This proof is short and uses two theorems you already know. The Extreme Value Theorem gives a smallest value m and a largest value M, because f is continuous on a closed interval. Every height is between them, so the area is between the area of a short rectangle and the area of a tall one.

Dividing by the width, which is positive, keeps the inequalities pointing the same way. So the average value is a number somewhere between m and M.

The last step is the Intermediate Value Theorem: a continuous function that takes the values m and M takes every value in between, including the average. That is the only place continuity is used a second time, and it is exactly the step that fails for a function with a jump.

11. Example 5.15: the average of a line

Worked example

Find the average value of the function below on the interval from 0 to 4, and find c where f equals that average.

\[ f(x) = 8 - 2x, \quad [0, 4] \]

Write the average as an integral

Why: Width 4 minus 0.

\[ f_{\text{ave}} = \frac{1}{4-0}\int_0^4 (8-2x)\,dx \]

Find the area geometrically

Why: The line meets the axes at 8 and 4: a right triangle.

\[ \int_0^4 (8-2x)\,dx = \tfrac12(4)(8) = 16 \]

Divide by the width

Why: Area over base is the average height.

\[ f_{\text{ave}} = \tfrac14(16) = 4 \]

Set f(c) equal to the average

Why: Solve the linear equation.

\[ 8 - 2c = 4 \;\Longrightarrow\; c = 2 \]

Figure (svg): The line y equals 8 minus 2x from x equals 0 to 4, forming a right triangle with the axes; a rectangle of height 4 over the same base has the same area, and the line crosses its top at c equals 2.

The triangle's area is 16 and the base is 4, so the average height is 4. The line reaches height 4 exactly halfway, at c equals 2.

Check

Why: Substitute c back in, and check that c lies in the interval.

\[ f(2) = 8 - 4 = 4 \;\checkmark, \qquad 2 \in [0,4] \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 474 — Example 5.15 and Figure 5.26

Here you do not need any calculus to find the integral, because the region is a triangle. The line meets the y-axis at 8 and the x-axis at 4, so the area is one half times 4 times 8, which is 16.

Dividing by the width 4 gives an average of 4. Then finding c is just solving a linear equation: 8 minus 2c equals 4, so c equals 2.

Look at the figure. The line crosses the top of the rectangle exactly in the middle of the interval. That is no accident: a line is symmetric about its midpoint, so it is above its average for half the interval and below it for the other half. Keep that fact in mind, because it is exactly what fails for curves.

12. Checkpoint 5.14: the average of x over 2

Worked example

\[ f(x) = \frac{x}{2}, \quad [0, 6] \]

Find the area

Why: Another triangle: base 6, height f(6), which is 3.

\[ \int_0^6 \frac{x}{2}\,dx = \tfrac12(6)(3) = 9 \]

Divide by the width

Why: The interval has width 6.

\[ f_{\text{ave}} = \frac{9}{6} = \frac32 \]

Solve f(c) equal to three halves

Why: Multiply both sides by 2.

\[ \frac{c}{2} = \frac32 \;\Longrightarrow\; c = 3 \]

Check

Why: For any line, the average is the height at the midpoint, and 3 is the midpoint of the interval from 0 to 6.

\[ f(3) = \tfrac32, \qquad \tfrac{0+6}{2} = 3 \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 474 — Checkpoint 5.14

Try this one yourself before stepping through it. The region under x over 2 from 0 to 6 is again a triangle, with base 6 and height 3, so its area is 9.

Dividing 9 by the width 6 gives an average of three halves. Setting c over 2 equal to three halves gives c equal to 3.

The check uses the fact from the previous slide: for any straight line, the average value on an interval is the value at the midpoint. The midpoint of the interval from 0 to 6 is 3, which matches. If your answer for a linear function is not the midpoint, something went wrong.

13. Example 5.16: where x squared equals its average

Worked example

You are given the integral below. Find c where the function equals its average on the interval from 0 to 3.

\[ \int_0^3 x^2\,dx = 9, \qquad f(x) = x^2 \]

Compute the average

Why: Divide the given area by the width 3.

\[ f_{\text{ave}} = \frac{1}{3-0}\int_0^3 x^2\,dx = \frac13(9) = 3 \]

Set f(c) equal to 3

Why: Replace f of c by c squared.

\[ c^2 = 3 \;\Longrightarrow\; c = \pm\sqrt3 \]

Keep the root inside the interval

Why: Minus root 3 is not in the interval from 0 to 3.

\[ c = \sqrt3 \approx 1.732 \]

Figure (svg): The parabola y equals x squared from 0 to 3, shaded, with a rectangle of height 3 over the same base. The curve crosses the top of the rectangle at x equals root 3, about 1.73, to the right of the midpoint 1.5, where the curve is only 2.25 high.

The parabola is steeper on the right, so it catches up with its average to the right of the midpoint, at root 3.

Check

Why: Square it back, and compare with the midpoint 1.5, where the curve is too low.

\[ (\sqrt3)^2 = 3 \;\checkmark, \qquad (1.5)^2 = 2.25 \ne 3 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 474-475 — Example 5.16 and Figure 5.27

The book hands you the integral here, 9, because you do not yet have a quick way to compute it. By the end of this lesson you will: it is 3 cubed over 3.

Dividing by the width 3 gives an average of 3. Solving c squared equal to 3 gives two roots, plus and minus root 3, but only the positive one lies in the interval from 0 to 3. Always check which solutions of the equation are actually in the interval; the theorem only speaks about points inside it.

Now look at the figure. The crossing point, root 3, is to the right of the midpoint 1.5. The parabola is flat on the left and steep on the right, so it spends a long time below its average and then shoots up past it. At the midpoint it has only reached 2.25.

14. Checkpoint 5.15: find the average, then c

Fill the middle

You are given the integral from 0 to 3 of two x squared minus 1, which is 15. Fill in the average and the value of c in the interval.

Fill in the blanks

f_5} = \frac5\sqrt{3} = ___, \quad 2c^2 - 1 = ___ \;\Longrightarrow\; c = ___

Why: Fifteen over a width of 3 is an average of 5. Then 2c squared minus 1 equals 5 gives c squared equal to 3, and only the positive root, root 3, lies in the interval from 0 to 3.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 475 — Checkpoint 5.15

Fill in the blanks in order. The first is the average: the given integral, 15, divided by the width of the interval, 3.

The second blank is the same number, because finding c means setting the function equal to the average. Then solve for c: add 1 to both sides, divide by 2, and take the square root.

You should get root 3 again, the same point as in Example 5.16. That is not a coincidence. Two x squared minus 1 is just x squared stretched and shifted, and stretching and shifting move the average in exactly the same way they move every value, so the crossing point stays put. Type root 3 as sqrt(3) or as a decimal.

15. Left of the midpoint, or right?

Prediction

\[ f(x) = x^3 \text{ on } [0, 2], \qquad f_{\text{ave}} = \frac12\int_0^2 x^3\,dx = 2 \]

Predict first

Before solving, where is the point c with c cubed equal to 2: left of the midpoint 1, at it, or right of it?

  • Left of the midpoint
  • Exactly at the midpoint
  • Right of the midpoint
  • There is no such c

Correct: Right of the midpoint

Why: At the midpoint the cube is only 1, below the average 2, because a curve that bends upward spends most of the interval low and makes up the area at the right end. So c is further right: the cube root of 2, about 1.26.

\[ c^3 = 2 \;\Longrightarrow\; c = \sqrt[3]{2} \approx 1.26 > 1 \]

Make your prediction before you calculate. The average of x cubed on the interval from 0 to 2 is 2, and you want to know where the cube reaches 2.

At the midpoint, x equal to 1, the cube is only 1, well below the average. A curve that bends upward like this one is low for most of the interval and makes up its area in a rush at the right end, so the point where it reaches its average is pushed to the right.

Solving confirms it: c is the cube root of 2, about 1.26. The general lesson is that the midpoint is the right answer for lines only. For a curve that bends up, expect c to the right of the midpoint; for a curve that bends down, to the left.

16. Break the theorem: drop continuity

Counterexample

Figure (svg): A step function equal to 0 on the interval from 0 to 1 and equal to 2 from 1 to 2, with a dashed line at its average value 1. The graph never touches the dashed line.

The average height is 1, but the function jumps straight from 0 to 2. Without continuity, nothing forces it through the value in between.

Discussion prompt

Give a function on the interval from 0 to 2 whose average value is never one of its values. Which hypothesis of Theorem 5.3 does it violate, and which step of the proof breaks?

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 473 — the hypotheses of Theorem 5.3

Look at the figure. The function is 0 on the left half of the interval and 2 on the right half. The two pieces have equal width, so the average height is 1, and yet the graph never touches the dashed line at height 1.

Go back to the proof and find the step that breaks. The squeeze still works: the average is between the minimum 0 and the maximum 2. What fails is the Intermediate Value Theorem, because the function jumps from 0 to 2 at x equal to 1 without passing through anything in between.

Counterexamples like this are how you learn why a hypothesis is in a theorem. Whenever you use the Mean Value Theorem for Integrals, check continuity first, and remember this picture of a graph leaping over its own average.

17. Trap: c is the midpoint

Trap

The trap

For x squared on the interval from 0 to 3:

\[ c = \frac{0 + 3}{2} = 1.5 \]

Wrong. The midpoint is where a line hits its average, not a curve.

The fix

Solve f(c) equal to the average. Here the average is 3, and the parabola reaches 3 at root 3. At the midpoint it is only 2.25.

\[ c^2 = 3 \Rightarrow c = \sqrt3 \]

\[ (1.5)^2 = 2.25 \ne 3 \]

The midpoint guess is tempting because it is right for every example with a straight line, and the first examples you meet are often lines. It is also right for any function that is symmetric about the midpoint of the interval.

For anything else you must actually solve f of c equal to the average. With x squared on the interval from 0 to 3, the average is 3, the midpoint gives only 2.25, and the true answer is root 3, about 1.73. Being wrong by a quarter of a unit here would cost you the whole question.

18. Part 1: the area function

Section

Part 2

19. An integral with a moving end is a function

Concept

Fix the lower limit at a and let the upper limit be a variable x. For each x the integral is a number, so the rule that sends x to that number is a function.

\[ F(x) = \int_a^x f(t)\,dt \]

The letter t is the dummy variable that runs along the interval; x only marks where the interval stops. For f(t) equal to 2 minus t, starting at 0:

\[ F(x) = \int_0^x (2-t)\,dt = 2x - \frac{x^2}{2} \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 475-476 — equation 5.16 and the remarks on notation

This is the idea that confuses people most in this section, so slow down. A definite integral is a number. But if you leave the upper limit as a letter, then for each choice of that letter you get a number, and a rule that turns inputs into numbers is a function.

That is why the integrand is written with t instead of x. The letter t is the variable that runs along the interval, from a up to x, while x says where to stop. Writing x in both places would make the expression ambiguous, which is why it is never done.

For the simple integrand 2 minus t you can find the area function explicitly by geometry or by antidifferentiating: it is 2x minus x squared over 2. Notice that its derivative is 2 minus x, the integrand with x put in place of t. That is Part 1 of the theorem, happening in front of you.

20. The sign of f steers the graph of F

Picture it

Figure (svg): Two panels. Left: the line f of t equals 2 minus t, with area shaded from t equals 0 to t equals 3: a positive triangle above the axis up to t equals 2 and a negative piece below it from 2 to 3. Right: the area function F of x equals 2x minus x squared over 2, rising to its peak 2 at x equals 2 and falling to 1.5 at x equals 3.

Each value of F is one whole shaded area. F climbs while f is positive, peaks exactly where f crosses zero, and falls once f turns negative.

Every point on the right-hand graph is a whole shaded area on the left. While f is positive the area grows; where f crosses zero the area stops growing; once f is negative the area shrinks. That is exactly how a function behaves when f is its derivative.

Read the two panels together. On the left, the shaded area stops at t equal to 3. On the right, the dot at height 1.5 records that whole area as a single number: the positive triangle of area 2, minus the negative sliver of area one half.

Now move the stopping point mentally from 0 to 4. While the line is above the axis, each step adds area and the right-hand graph climbs. At t equal to 2 the line crosses zero, nothing is being added, and the graph of F is at its peak. After that the line is below the axis, area is being subtracted, and the graph falls.

That is exactly the relationship between a function and its derivative from Chapter 4: F increases where f is positive, has a maximum where f changes from positive to negative, and decreases where f is negative. The picture is already telling you that f is the derivative of F.

21. Nudge the right edge and watch the area grow

Picture it

Figure (svg): A curve with the region under it shaded from a to x, labelled F of x. Next to it a thin extra strip from x to x plus h is highlighted; its height is about f of x and its width is h.

Nudging x forward by h adds one thin strip. Its area is nearly height times width, f of x times h, so the rate at which the area grows is f of x.

Move the right end from x to x plus h. The area gains one thin strip, and the strip is almost a rectangle with height f(x) and width h.

\[ F(x+h) - F(x) = \int_x^{x+h} f(t)\,dt \approx f(x)\,h \]

Focus on the thin yellow strip. It is the extra area you gain when the right edge moves from x to x plus h. Its width is h and its height is very nearly f of x, the pink segment, so its area is very nearly f of x times h.

Divide by h and you get the average rate at which the area grew over that small step: very nearly f of x. As h shrinks, the strip becomes a better and better rectangle, and in the limit the rate is exactly f of x.

That is the whole of Part 1 in one picture: the rate of change of area is height. The formal proof on the next slides only makes the word nearly precise, and the Mean Value Theorem for Integrals is the tool that does it.

22. Theorem 5.4: the Fundamental Theorem, Part 1

Concept

Divide the strip by its width and let the width shrink. The approximation becomes exact, and it says the area function's rate of change is the height of the curve.

Fundamental Theorem of Calculus, Part 1 (Theorem 5.4) — If f is continuous on the interval from a to b and F(x) is the integral of f from a to x, then F is differentiable and its derivative is f, for every x strictly between a and b.

\[ \frac{d}{dx}\int_a^x f(t)\,dt = f(x) \]

Integrate, then differentiate, and you get back where you started. In particular every continuous function has an antiderivative: its own area function.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 475-476 — Theorem 5.4

Read the statement slowly. You start with a continuous function f, build its area function by integrating from a fixed point a up to x, and then differentiate. What comes out is f again.

In practice this means that to differentiate an integral whose upper limit is x, you do no integrating at all. You simply write down the integrand with x in place of the dummy variable.

There is a deep consequence in the last sentence. Every continuous function has an antiderivative, namely its own area function, even when you cannot write that antiderivative with familiar functions. Part 1 is where the existence of antiderivatives comes from.

23. Reading Theorem 5.4 piece by piece

Notation

Annotate

On: \( F(x) = \int_a^x f(t)\,dt \;\Longrightarrow\; F'(x) = f(x) \)

  • A fixed starting point. Changing it adds a constant to F, which the derivative ignores.
  • The only place x appears on the left. It is the moving right edge.
  • t is a dummy variable: rename it to s or u and nothing changes. Never write x inside as well.
  • Put x where t was. No integrating at all is needed.

Step through the notes. The lower limit a is the fixed starting point, and it does not affect the answer: starting somewhere else only adds or removes a fixed chunk of area, a constant, and constants differentiate to zero.

The upper limit x is the only place x appears on the left side. Inside the integral the variable is t, a dummy you could rename freely. The answer is obtained by putting x where t was.

A common slip is to write the integrand in terms of x from the start, as in the integral from a to x of f of x dx. It looks harmless and it is meaningless, because the same letter is then asked to be both the moving point and the stopping point.

24. The proof of Part 1, line by line

Concept

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

\[ = \lim_{h\to 0}\frac1h\left[\int_a^{x+h} f(t)\,dt - \int_a^{x} f(t)\,dt\right] \]

\[ = \lim_{h\to 0}\frac1h\int_x^{x+h} f(t)\,dt \]

That last fraction is the average of f over the short interval from x to x plus h. By the Mean Value Theorem for Integrals it equals f(c) for some c in that interval.

\[ = \lim_{h\to 0} f(c), \qquad x \le c \le x + h \]

As h shrinks, c is squeezed onto x, and f is continuous, so f(c) tends to f(x).

\[ F'(x) = f(x) \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 476 — proof of Theorem 5.4

Follow the lines one at a time. The first is just the definition of the derivative. The second writes F at x plus h and F at x as integrals. The third uses the fact that the integral from a to x plus h, minus the integral from a to x, leaves the integral from x to x plus h: the strip from the picture.

The fourth line is the key move. One over h times the integral over an interval of width h is precisely the average value of f on that interval. The Mean Value Theorem for Integrals says that average is actually a value of f, at some point c trapped between x and x plus h.

Finally, as h goes to zero, the trap closes and c is forced onto x. Because f is continuous, f of c tends to f of x. Notice that continuity was used twice: once inside the Mean Value Theorem for Integrals and once in this last limit.

25. Slide the lower limit and watch what changes

Tweak it

Parameter explorer

The curve is the area function of f(t) = t, starting from a. Slide a. What happens to the graph, and what happens to its slope at any given x?

\[ \int_{{a}}^{x} t\,dt \]

  • a — from -2 to 2: lower limit a

\[ \int_a^x t\,dt = \frac{x^2}{2} - \frac{a^2}{2} \;\Longrightarrow\; F'(x) = x \]

Move the slider for a from minus 2 to 2 and watch the curve. It keeps exactly the same shape and only slides up and down. Starting the area from a different point adds or removes a fixed amount of area, the same amount for every x.

Now think about slopes. Sliding a graph up or down does not change its slope anywhere. At any x the slope is x, the value of the integrand there, no matter where the area starts.

This is why Part 1 does not mention a in its answer, and why the family of area functions from different starting points is a family of antiderivatives differing by constants. It is the same plus C you met in Section 4.10, now with a picture.

26. Example 5.17: a derivative with no integrating

Worked example

Use Part 1 to find the derivative of the function below.

\[ g(x) = \int_1^x \frac{1}{t^3+1}\,dt \]

Check the hypotheses

Why: The integrand is continuous for t at least 1, since the denominator is never zero there, and the upper limit is plain x.

\[ f(t) = \frac{1}{t^3+1} \]

Apply Part 1

Why: Replace t by x in the integrand.

\[ g'(x) = \frac{1}{x^3+1} \]

Check with a difference quotient

Why: Computing g numerically near x equal to 1 and 2 and taking symmetric difference quotients gives the same slopes.

\[ g'(1) \approx 0.500000 = \tfrac12, \qquad g'(2) \approx 0.111111 = \tfrac19 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 476 — Example 5.17

This integrand, one over t cubed plus 1, does have an antiderivative, but it is long and messy, with a logarithm and an arctangent. You do not need it. The question asks for the derivative of the integral, and Part 1 gives that directly.

Check the two conditions first: the upper limit is plain x, and the integrand is continuous on the interval, since t cubed plus 1 is at least 2 for t at least 1. Then put x in place of t.

The check is worth understanding. The numbers on the slide come from computing g numerically, by a very fine Riemann-type sum, at points just either side of 1 and 2, and dividing the change by the step. The slopes agree with one half and one ninth to six decimal places.

27. Checkpoint 5.16: a different letter outside

Prediction

\[ g(r) = \int_0^r \sqrt{x^2+4}\,dx \]

Predict first

What is g′(r)?

  • √(r² + 4)
  • √(x² + 4)
  • √(r² + 4) − 2
  • r/√(r² + 4)

Correct: √(r² + 4)

Why: Here the dummy variable is x and the moving limit is r, so Part 1 says: put r where x was. The answer must be a function of r alone. Subtracting the value at 0 would be Part 2 thinking, and r over the root is what you get by differentiating the integrand instead of the integral. Numerically, at r = 1 a difference quotient gives 2.2361, which is root 5.

\[ g'(r) = \sqrt{r^2+4}, \qquad g'(1) = \sqrt5 \approx 2.2361 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 477 — Checkpoint 5.16

The letters have been reshuffled on purpose. Here x is the dummy variable inside the integral and r is the moving upper limit. Do not let the familiar letter x fool you into thinking it is the variable you differentiate with respect to.

The rule is unchanged: put the upper limit where the dummy variable was. So the derivative is the square root of r squared plus 4, a function of r.

Each wrong option is a real mistake. Writing the answer in x confuses the dummy with the limit. Subtracting 2 subtracts the integrand at the lower limit, which is Part 2 thinking in a Part 1 question. And r over the root is what you get if you differentiate the integrand, which is not what was asked.

28. Part 1 hands you antiderivatives you cannot write down

Concept

Exercise 148. No combination of powers, roots, exponentials, logarithms or trig functions has derivative e to the minus t squared. Part 1 does not care.

\[ \frac{d}{dx}\int_1^x e^{-t^2}\,dt = e^{-x^2} \]

So the area function below is a perfectly good antiderivative, defined by its integral. Up to a constant factor it is the error function behind every bell-curve probability.

\[ E(x) = \int_1^x e^{-t^2}\,dt, \qquad E'(1.5) = e^{-2.25} \approx 0.1054 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 484 — Exercise 148

This is the most surprising consequence of Part 1. Some continuous functions, e to the minus t squared among them, have antiderivatives that cannot be written as any combination of the functions you know. This has been proved; it is not that nobody has found the formula yet.

Yet Part 1 tells you the derivative of its area function instantly: e to the minus x squared. The area function is a genuine antiderivative, defined by its integral, and you can compute its values as accurately as you like with Riemann sums.

This is not a curiosity. The standard normal distribution in statistics is built from exactly this area function, and your calculator's normal probability key is evaluating it numerically. Defining a new function as an integral is a normal, respectable thing to do.

29. Find the error: subtracting the bottom end

Error analysis

Annotate

On: \( \frac{d}{dx}\int_2^x (t^3+1)\,dt = (x^3 + 1) - 9 \)

  • That is f(2), the integrand at the lower limit. It was subtracted as if this were Part 2, F(b) − F(a). But F(a) is a constant, and its derivative is zero.
  • Part 1 gives just x³ + 1. The lower limit never appears in the derivative.
  • The integral is x⁴/4 + x − 6; its derivative is x³ + 1, with no − 9 anywhere.

Step through the notes. The minus 9 is the integrand evaluated at the lower limit 2, since 2 cubed plus 1 is 9. The writer has mixed up two different operations: differentiating an integral with Part 1, and evaluating one with Part 2.

Think about what the lower limit does. It fixes where the area starts, which contributes a constant to F. The derivative of a constant is zero, so the lower limit leaves no trace in the answer. The correct derivative is simply x cubed plus 1.

The check in the last note settles it: evaluate the integral explicitly and differentiate. The constant from the lower limit, minus 6, disappears on differentiation. Whenever you are unsure, this evaluate-then-differentiate check is available for any integrand you can antidifferentiate.

30. A number, or a function of x?

Sorting

Sort into buckets

Sort each expression: is it a single number, or a function of x?

A number
∫ from 0 to 1 of t² dt; ∫ from −2 to 2 of (t² − 4) dt; d/dx of ∫ from 0 to 1 of t² dt
A function of x
∫ from 0 to x of t² dt; ∫ from x to 5 of eᵗ dt; ∫ from 1 to x of 1/(t³ + 1) dt
num
Both limits are numbers, so the integral is one fixed number (1/3 and −32/3 here). The derivative of a fixed number is 0, which is also a number.
fn
x sits in a limit of integration, so the value depends on x. The letter inside, t, is only a dummy.

Sort each expression by asking one question: does the letter x appear in a limit of integration? If it does, the value changes as x changes, so the expression is a function of x. If both limits are numbers, the expression is one fixed number, however complicated the integrand.

The letter inside the integral, t, never matters for this question. It is a dummy that disappears once the integral is evaluated.

The derivative of the integral from 0 to 1 is the one people get wrong. That integral is the number one third, and the derivative of any number is zero. Part 1 does not apply to it, because its upper limit is not a variable.

31. Variable limits and the chain rule

Section

Part 3

32. When the upper limit is a function of x

Concept

If the upper limit is u(x), the integral is the area function G evaluated at u(x): a composition. Differentiate it with the chain rule.

\[ G(u) = \int_a^u f(t)\,dt, \qquad G'(u) = f(u) \]

\[ \frac{d}{dx}\int_a^{u(x)} f(t)\,dt = \frac{d}{dx}G\big(u(x)\big) = f\big(u(x)\big)\,u'(x) \]

Two moves: put the limit into the integrand, then multiply by the limit's derivative. When u(x) is just x, the second factor is 1 and this is Part 1 again.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 477 — Example 5.18 solution

Read the integral as a composition. The inner function takes x to u of x. The outer function G takes a stopping point u to the area up to u. Composing them gives the integral with u of x on top.

Part 1 tells you the derivative of the outer function: G prime of u is f of u. The chain rule from Chapter 3 then says: outer derivative evaluated at the inner function, times the derivative of the inner function.

So the recipe has two moves, and the second is the one people forget. First, put the upper limit into the integrand. Second, multiply by the derivative of that upper limit. When the limit is plain x, its derivative is 1 and the second move is invisible, which is why it is so easy to forget when it matters.

33. Example 5.18: the square root in the limit

Worked example

\[ F(x) = \int_1^{\sqrt x} \sin t\,dt, \qquad \text{find } F'(x) \]

Name the inner function

Why: The upper limit is the function u of x.

\[ u(x) = \sqrt x, \quad u'(x) = \frac{1}{2\sqrt x} \]

Put the limit into the integrand

Why: Part 1 on the outer layer.

\[ F'(x) = \sin\big(u(x)\big)\cdot u'(x) \]

Multiply by the limit's derivative

Why: Substitute both pieces.

\[ F'(x) = \sin\sqrt x \cdot \frac{1}{2\sqrt x} = \frac{\sin\sqrt x}{2\sqrt x} \]

Check by evaluating first

Why: This integrand has an antiderivative, so find F explicitly and differentiate it.

\[ F(x) = \big[-\cos t\big]_1^{\sqrt x} = -\cos\sqrt x + \cos 1 \]

Check, continued

Why: Differentiate: the derivative of minus cosine is sine, times the chain factor. It matches.

\[ F'(x) = \sin\sqrt x\cdot\frac{1}{2\sqrt x} \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 477 — Example 5.18

Start by naming the inner function. The upper limit is the square root of x, so its derivative is one over twice the square root of x.

Then do the two moves. Put root x into the integrand, giving sine of root x. Multiply by the derivative of root x. The result is sine of root x over twice root x.

This integrand, sine, is one you can antidifferentiate, so the book's answer can be checked the long way. Evaluate the integral to get minus cosine of root x plus cosine of 1, then differentiate with the ordinary chain rule. The same answer appears, and the constant cosine of 1 vanishes, exactly as the lower limit always does.

34. Checkpoint 5.17: x cubed in the limit

Worked example

\[ F(x) = \int_1^{x^3} \cos t\,dt, \qquad \text{find } F'(x) \]

Name the inner function

Why: The upper limit and its derivative.

\[ u(x) = x^3, \quad u'(x) = 3x^2 \]

Apply Part 1 and the chain rule

Why: Put x cubed into cosine, then multiply by 3x squared.

\[ F'(x) = \cos(x^3)\cdot 3x^2 = 3x^2\cos(x^3) \]

Check by evaluating first

Why: The antiderivative of cosine is sine.

\[ F(x) = \sin(x^3) - \sin 1 \;\Longrightarrow\; F'(x) = 3x^2\cos(x^3) \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 477 — Checkpoint 5.17

Try this before stepping through it. The inner function is x cubed, with derivative 3x squared.

Put x cubed into the cosine and multiply by 3x squared. Write the polynomial factor in front, 3x squared times cosine of x cubed, so that nobody reads it as the cosine of something bigger.

The check evaluates first: the antiderivative of cosine is sine, so the integral is the sine of x cubed minus the sine of 1. Differentiating that with the chain rule gives the same answer. Both routes must agree, and when the integrand is one you can antidifferentiate, doing both is the surest way to catch a slip.

35. Trap: forgetting the chain-rule factor

Trap

The trap

Plugging the limit in and stopping:

\[ \frac{d}{dx}\int_0^{x^2}\cos t\,dt \]

\[ = \cos(x^2) \]

Wrong. The limit is x squared, not x.

The fix

The integral is a function of x squared, so the chain rule multiplies by the derivative of x squared.

\[ = \cos(x^2)\cdot 2x \]

Check: the integral is the sine of x squared, whose derivative carries the same 2x.

This slip is common because Part 1 with a plain x on top trains you to plug in and stop. The moment the limit is anything other than x, there is a second factor to include.

A quick way to catch it: ask what the integral actually is. Here it is the sine of x squared, and you already know that differentiating the sine of x squared produces a factor of 2x. The answer to a Part 1 question must agree with the ordinary chain rule, because it is the ordinary chain rule.

36. A variable lower limit brings a minus sign

Concept

Swapping the limits of an integral changes its sign. Use that to move the variable to the top.

\[ \int_{x}^{b} f(t)\,dt = -\int_{b}^{x} f(t)\,dt \]

\[ \frac{d}{dx}\int_x^b f(t)\,dt = -f(x) \]

When both limits move, split at any fixed number and apply the rule to each piece: top limit plus, bottom limit minus.

\[ \frac{d}{dx}\int_{v(x)}^{u(x)} f(t)\,dt = f\big(u(x)\big)u'(x) - f\big(v(x)\big)v'(x) \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 477 — Example 5.19 solution

Reversing the limits of an integral changes its sign, a property from Section 5.2. So an integral with x at the bottom is minus an integral with x at the top, and Part 1 then gives minus f of x.

The general formula on the slide looks heavy but is one idea: every limit that depends on x contributes its own term. The top limit contributes plus the integrand at the top times the top's derivative. The bottom limit contributes minus the integrand at the bottom times the bottom's derivative.

It helps to think about area. Moving the right edge to the right adds area, so the top limit's term is positive. Moving the left edge to the right removes area, so the bottom limit's term is negative.

37. Example 5.19: both limits move

Worked example

\[ F(x) = \int_x^{2x} t^3\,dt, \qquad \text{find } F'(x) \]

Split at 0

Why: Both limits depend on x, so break the integral at a fixed point.

\[ F(x) = \int_x^{0} t^3\,dt + \int_0^{2x} t^3\,dt \]

Flip the first piece

Why: Reversing the limits changes the sign.

\[ F(x) = -\int_0^{x} t^3\,dt + \int_0^{2x} t^3\,dt \]

Differentiate the first piece

Why: Plain Part 1, with the minus sign carried along.

\[ \frac{d}{dx}\left[-\int_0^{x} t^3\,dt\right] = -x^3 \]

Differentiate the second piece

Why: Chain rule with u equal to 2x.

\[ \frac{d}{dx}\int_0^{2x} t^3\,dt = (2x)^3\cdot 2 = 16x^3 \]

Add

Why: Combine like terms.

\[ F'(x) = -x^3 + 16x^3 = 15x^3 \]

Figure (svg): The curve y equals t cubed from 0 to 2.3. The area from 0 to x equals 1 is shaded red and the area from 1 to 2 is shaded green: the integral from x to 2x is the big area from 0 to 2x minus the small one from 0 to x.

Both ends move. Split at a fixed point, here 0: the area from x to 2x is the area up to 2x minus the area up to x.

Check by evaluating first

Why: Integrate directly, then differentiate.

\[ F(x) = \left[\frac{t^4}{4}\right]_x^{2x} = \frac{16x^4 - x^4}{4} = \frac{15x^4}{4} \;\Longrightarrow\; F'(x) = 15x^3 \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 477-478 — Example 5.19

When both limits depend on x, Part 1 does not apply directly, because it needs a fixed lower limit. The fix is to split the integral at any fixed number; 0 is convenient here. Now each piece has one constant limit and one moving limit.

Flip the first piece so that x is on top, and remember the minus sign that comes with the flip. Differentiating that piece gives minus x cubed. The second piece has 2x on top, so it gives 2x cubed, which is 8x cubed, times the derivative 2: sixteen x cubed.

Look at the figure: the area from x to 2x is the big area up to 2x minus the small one up to x, which is exactly the split. The check evaluates the integral directly, fifteen x to the fourth over 4, and differentiates it to get fifteen x cubed again.

38. Checkpoint 5.18: from x to x squared

Worked example

\[ F(x) = \int_x^{x^2} \cos t\,dt, \qquad \text{find } F'(x) \]

Top limit: plug in and multiply

Why: u equals x squared, u prime equals 2x.

\[ \cos(x^2)\cdot 2x \]

Bottom limit: plug in, multiply, subtract

Why: v equals x, v prime equals 1.

\[ \cos(x)\cdot 1 \]

Combine

Why: Top minus bottom.

\[ F'(x) = 2x\cos(x^2) - \cos x \]

Check by evaluating first

Why: F is the sine of x squared minus the sine of x. At x equal to 2, a numerical difference quotient gives the same number.

\[ F(x) = \sin(x^2) - \sin x, \qquad F'(2) = 4\cos 4 - \cos 2 \approx -2.1984 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 478 — Checkpoint 5.18

Use the top-minus-bottom rule directly this time. The top limit is x squared, so its term is the cosine of x squared times 2x. The bottom limit is x, so its term is the cosine of x times 1, and it is subtracted.

Putting them together gives 2x times the cosine of x squared, minus the cosine of x. Writing the answer out in this order, top term first, helps you keep the sign straight.

For the check, evaluate first: the antiderivative of cosine is sine, so the integral is the sine of x squared minus the sine of x, and differentiating that gives the same expression. At x equal to 2 both the formula and a numerical difference quotient give about minus 2.198, a negative slope you would not have guessed.

39. Exercise 153: a derivative that comes out constant

Worked example

\[ \frac{d}{dx}\int_0^{\sqrt x} t\,dt \]

Name the inner function

Why: The upper limit is the square root.

\[ u(x) = \sqrt x, \quad u'(x) = \frac{1}{2\sqrt x} \]

Plug in and multiply

Why: The integrand is t, so f(u) is just u.

\[ \sqrt x\cdot\frac{1}{2\sqrt x} \]

Simplify

Why: The roots cancel.

\[ = \frac12 \]

Check by evaluating first

Why: The integral is half of t squared, from 0 to root x, which is x over 2, a line of slope one half.

\[ \int_0^{\sqrt x} t\,dt = \frac{(\sqrt x)^2}{2} = \frac{x}{2} \;\Longrightarrow\; \frac{d}{dx}\frac{x}{2} = \frac12 \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 484 — Exercise 153

This exercise from the section's set looks like it should give something complicated, and it gives one half. Work it with the chain-rule recipe: the limit is root x, the integrand at that limit is root x, and the derivative of the limit is one over twice root x.

The roots cancel, leaving one half. When a result is that simple, it is a sign that something structural is going on, and the check reveals it.

Evaluate the integral first: the antiderivative of t is t squared over 2, and at root x that is x over 2. So the whole integral is just the line x over 2, whose slope is one half everywhere. Two different routes, the same answer, and the second one explains why.

40. Match each integral to its derivative

Matching

Match the pairs

  • a. d/dx ∫ from 1 to x of cos t dt
  • b. d/dx ∫ from 1 to x² of cos t dt
  • c. d/dx ∫ from x to 1 of cos t dt
  • d. d/dx ∫ from x to x² of cos t dt
  • w. cos x
  • x. 2x cos(x²)
  • y. −cos x
  • z. 2x cos(x²) − cos x

Why: One rule covers all four: the top limit contributes f(top) times its derivative, the bottom limit contributes minus f(bottom) times its derivative, and a constant limit contributes nothing.

All four integrals have the same integrand, cosine of t. Only the limits change, and the limits alone decide the answer.

Work through them with one rule. A constant limit contributes nothing. A limit of x on top contributes cosine of x. A limit of x at the bottom contributes minus cosine of x. A limit of x squared contributes cosine of x squared times 2x, with a plus sign on top and a minus sign at the bottom.

If you mixed up the second and fourth, look again at the lower limits: in the second it is the constant 1, in the fourth it is x, and that extra moving end brings its own term.

41. Exercise 193: symmetric limits

Prediction

\[ \frac{d}{dx}\int_{-x}^{x}\sin t\,dt \]

Predict first

Using top-minus-bottom, what is this derivative?

  • 2 sin x
  • 0
  • −2 cos x
  • 2 cos x

Correct: 0

Why: The top gives sin x times 1. The bottom gives minus sin(−x) times (−1), which is minus sin x. They cancel. Check by evaluating: the integral is −cos x + cos(−x), and cosine is even, so the integral is 0 for every x. Sine is odd; its area on a symmetric interval always cancels.

\[ \sin x\cdot 1 - \sin(-x)\cdot(-1) = \sin x - \sin x = 0 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 485 — Exercise 193

Predict first, then work it through with top minus bottom. The top limit x contributes the sine of x times 1. The bottom limit minus x contributes minus the sine of minus x, times the derivative of minus x, which is minus 1.

Sine is odd, so the sine of minus x is minus the sine of x. The bottom term is therefore minus the sine of x, and the two terms cancel to zero.

The picture explains it. Sine is odd, so its graph on the left of zero is the upside-down copy of its graph on the right. On any interval symmetric about zero, the positive area and the negative area are equal and cancel. The integral is zero for every x, so its derivative is zero too.

42. Part 2: evaluating integrals

Section

Part 4

43. Theorem 5.5: the evaluation theorem

Concept

Fundamental Theorem of Calculus, Part 2 (Theorem 5.5) — If f is continuous on the interval from a to b and F is any antiderivative of f there, then the definite integral of f from a to b equals F(b) minus F(a).

\[ \int_a^b f(x)\,dx = F(b) - F(a) = F(x)\Big|_a^b \]

The vertical bar is shorthand: evaluate at the top, evaluate at the bottom, subtract. The limit of Riemann sums from the start of the lesson is now two lines.

\[ \int_0^2 x^2\,dx = \frac{x^3}{3}\Big|_0^2 = \frac83 - 0 = \frac83 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 478 — Theorem 5.5, equation 5.17

This is the part of the theorem you will use most often, and it is almost too good. To evaluate a definite integral, find any antiderivative, evaluate it at the upper limit, evaluate it at the lower limit, and subtract.

The vertical bar with limits is just shorthand for that last step. You will see it written with a bar, or with square brackets and limits; both mean top minus bottom.

Look at the last line. It is the integral from the start of the lesson, the one that took a sum-of-squares formula and a limit. Now it is x cubed over 3, evaluated at 2 and at 0. Same answer, eight thirds, and no limits in sight. The hard work of Riemann sums has been moved into the theorem, which you only have to prove once.

44. Reading Theorem 5.5 piece by piece

Notation

Annotate

On: \( \int_a^b f(x)\,dx = F(b) - F(a), \quad F' = f \)

  • A limit of Riemann sums: a signed area, one number.
  • Any antiderivative. You find it with the rules of Section 4.10, run backwards.
  • Top minus bottom, in that order. Swapping them flips the sign.
  • The hypothesis. If f blows up or jumps inside the interval, the theorem says nothing.

Step through the notes and notice what each side of the equation is. The left side is defined by Riemann sums: a signed area. The right side is defined by derivatives: a change in an antiderivative. The theorem claims they are equal, and that is a genuinely surprising claim.

The order in F of b minus F of a matters. Top minus bottom. If you subtract the other way you get exactly the negative of the correct answer, and a sign check will often catch it.

Do not skip the last note. The theorem needs f continuous on the whole closed interval. It is easy to forget, because most integrands you meet are continuous everywhere, and the trap at the end of this part shows what happens when they are not.

45. The proof of Part 2: the rises add up

Concept

Figure (svg): An increasing curve y equals F of x over five equal subintervals from a to b. At each partition point a short vertical arrow shows the rise of F over that subinterval; the five arrows stack up to the total rise F of b minus F of a, drawn as a long bracket on the right.

The total climb from a to b is the sum of the small climbs. The Mean Value Theorem turns each small climb into a rectangle's area.

Cut the interval into n pieces. The total change of F is the sum of its changes over the pieces, since everything in between cancels.

\[ F(b) - F(a) = \sum_{i=1}^{n}\big[F(x_i) - F(x_{i-1})\big] \]

On each piece the ordinary Mean Value Theorem finds a point where the slope of F, which is f, matches the secant slope.

\[ F(x_i) - F(x_{i-1}) = F'(c_i)\,\Delta x = f(c_i)\,\Delta x \]

\[ F(b) - F(a) = \sum_{i=1}^{n} f(c_i)\,\Delta x \;\xrightarrow{n\to\infty}\; \int_a^b f(x)\,dx \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 478-479 — proof of Theorem 5.5

Start with the figure. The total climb of F from a to b is the sum of the five small climbs, the yellow arrows. That is the first equation: in the sum, each intermediate value appears once with a plus and once with a minus, and everything cancels except the two ends.

Now apply the ordinary Mean Value Theorem from Section 4.4 to F on each small piece. It finds a point c i inside the piece where the tangent slope equals the secant slope. The tangent slope is F prime, which is f, so each small climb equals f of c i times the width of the piece: the area of a rectangle.

So F of b minus F of a equals a Riemann sum for f, exactly, for every n. Let n go to infinity: the right side becomes the integral, and the left side never changed. The proof never needed a formula for any sum.

46. Why the plus C never matters

Concept

Any two antiderivatives differ by a constant. Use F plus C instead of F and watch the constant cancel.

\[ \big[F(b) + C\big] - \big[F(a) + C\big] = F(b) - F(a) \]

So in a definite integral you may always take C equal to 0. The constant is not forgotten; it genuinely cancels, every time.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 479 — Analysis of Example 5.20

When you found an indefinite integral in Section 4.10, you always wrote plus C, because there is a whole family of antiderivatives. In a definite integral you can leave it out, and this slide shows why that is not carelessness.

Whatever constant you choose, it appears once in F of b and once in F of a, and the subtraction cancels it. Every antiderivative gives the same definite integral.

So in practice you pick the simplest antiderivative, the one with C equal to zero. The same fact appeared in the slider on lower limits: different starting points shift the area function by a constant, and constants disappear under subtraction just as they disappear under differentiation.

47. Example 5.20: a negative answer

Worked example

Use Part 2 to evaluate the integral below.

\[ \int_{-2}^{2}(t^2 - 4)\,dt \]

Find an antiderivative

Why: Power rule on each term; take C equal to 0.

\[ F(t) = \frac{t^3}{3} - 4t \]

Evaluate at both ends

Why: Top, then bottom.

\[ F(2) = \frac83 - 8, \qquad F(-2) = -\frac83 + 8 \]

Subtract

Why: Distribute the minus sign over the second bracket.

\[ \left(\frac83 - 8\right) - \left(-\frac83 + 8\right) = \frac83 - 8 + \frac83 - 8 \]

Simplify

Why: Collect the thirds and the eights.

\[ = \frac{16}{3} - 16 = -\frac{32}{3} \]

Figure (svg): The parabola f of t equals t squared minus 4, crossing the t-axis at minus 2 and 2, with the region between the curve and the axis on that interval shaded red, entirely below the axis.

The whole region sits below the axis, so the integral is negative. Its area is 32 over 3; the integral reports minus that.

Check with Archimedes

Why: A parabolic segment has area two thirds of its enclosing rectangle: here base 4, height 4. The region lies below the axis, so the integral is minus that area.

\[ \tfrac23(4)(4) = \tfrac{32}{3} \;\Longrightarrow\; \int_{-2}^{2}(t^2-4)\,dt = -\tfrac{32}{3} \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 479-480 — Example 5.20 and Figure 5.28

Work through the steps. The antiderivative of t squared minus 4 is t cubed over 3 minus 4t. Evaluate at 2, evaluate at minus 2, and subtract, taking care with the minus sign in front of the second bracket; this is where most errors in this example happen.

The answer, minus 32 over 3, is negative. Look at the figure to see why: on the whole interval the parabola is below the axis, so every Riemann sum rectangle has negative height, and the integral is a negative signed area.

The check is one Archimedes found two thousand years ago: the area of a parabolic segment is two thirds of the rectangle around it. Here that rectangle is 4 wide and 4 deep, so the area is 32 over 3, and the integral is minus that. Area is never negative; a definite integral can be.

48. Example 5.21: rewrite before integrating

Worked example

\[ \int_1^9 \frac{x-1}{\sqrt x}\,dx \]

Write the root as a power

Why: Radicals become rational exponents.

\[ \int_1^9 \frac{x-1}{x^{1/2}}\,dx \]

Split the numerator

Why: Each term over the denominator.

\[ \int_1^9 \left(\frac{x}{x^{1/2}} - \frac{1}{x^{1/2}}\right)dx \]

Simplify the powers

Why: Subtract exponents.

\[ \int_1^9 \left(x^{1/2} - x^{-1/2}\right)dx \]

Find an antiderivative

Why: Power rule: raise each exponent by one and divide by the new exponent.

\[ F(x) = \frac{2}{3}x^{3/2} - 2x^{1/2} \]

Evaluate at 9

Why: The square root of 9 is 3, so 9 to the three halves is 27.

\[ F(9) = \tfrac23(27) - 2(3) = 18 - 6 = 12 \]

Evaluate at 1 and subtract

Why: Every power of 1 is 1.

\[ F(1) = \tfrac23 - 2 = -\tfrac43, \qquad 12 - \left(-\tfrac43\right) = \tfrac{40}{3} \]

Figure (svg): The curve y equals x minus 1 over root x, crossing the axis at x equals 1 and rising slowly to about 2.67 at x equals 9, with the region under it from 1 to 9 shaded.

The region is 8 units wide and never taller than about 2.7, so an area a little over 13 is believable.

Check by differentiating F

Why: The derivative of the antiderivative must be the integrand; a Simpson's-rule estimate also gives 13.3333.

\[ F'(x) = x^{1/2} - x^{-1/2} = \frac{x-1}{\sqrt x} \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 480-481 — Example 5.21 and Figure 5.29

There is no quotient rule for antiderivatives, so the first job is to rewrite the integrand as a sum of powers. Replace the root with a one-half power, split the fraction into two terms, and subtract exponents. Now the power rule applies to each term.

Evaluating at 9 is easier than it looks if you remember that 9 to the one half is 3 and 9 to the three halves is 3 cubed, 27. Every power of 1 is 1, so the lower limit is quick. Then subtract, and watch the double negative: minus minus four thirds is plus four thirds.

The figure makes the size believable: a region 8 wide and at most about 2.7 tall has area a bit over 13. The check differentiates the antiderivative and gets back the rewritten integrand, which is exactly what Part 2 needs to be true.

49. Checkpoint 5.19: a negative power

Worked example

\[ \int_1^2 x^{-4}\,dx \]

Find an antiderivative

Why: Raise the exponent to minus 3 and divide by minus 3.

\[ F(x) = \frac{x^{-3}}{-3} = -\frac{1}{3x^3} \]

Evaluate at both ends

Why: Top, then bottom.

\[ F(2) = -\frac{1}{24}, \qquad F(1) = -\frac13 \]

Subtract

Why: Common denominator 24.

\[ -\frac{1}{24} + \frac{8}{24} = \frac{7}{24} \]

Check the size

Why: The integrand falls from 1 to one sixteenth over an interval of width 1, so the answer must lie between those. It does.

\[ \tfrac{1}{16} = 0.0625 < \tfrac{7}{24} \approx 0.2917 < 1 \;\checkmark \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 481 — Checkpoint 5.19

The power rule works for negative exponents too, as long as the exponent is not minus 1. Raise minus 4 to minus 3 and divide by minus 3; that gives minus one over 3x cubed.

At 2 the antiderivative is minus one over 24, and at 1 it is minus one third. Top minus bottom: minus one twenty-fourth plus one third, which with a common denominator is 7 over 24.

The size check is a habit worth building. On an interval of width 1 the integral is an average height, so it must lie between the smallest value of the integrand, one sixteenth, and the largest, 1. The answer, about 0.29, is comfortably inside that range.

50. Before integrating a quotient

Step zero

\[ \int_1^4 \frac{2 - \sqrt t}{t^2}\,dt \]

Discussion prompt

Exercise 180. There is no quotient rule for antiderivatives. What do you do before anything else, and what is the answer?

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 485 — Exercise 180

Answer the question before you look. The step zero here is recognising that the integrand is not ready for the power rule, because it is a quotient with a root in it.

Split the numerator over the denominator and turn every term into a single power of t. Two over t squared becomes 2 t to the minus 2, and root t over t squared becomes t to the minus three halves. After that, the integration is routine.

Evaluating the antiderivative at 4 gives one half; at 1 it gives zero. So the answer is one half. Rewriting is almost always the first step with fractions and roots, and skipping it is the most common reason people get stuck on problems like this.

51. Find the error: bottom minus top

Error analysis

Annotate

On: \( \int_1^3 2x\,dx = x^2\Big|_1^3 = 1 - 9 = -8 \)

  • The lower limit was evaluated first. Part 2 is F(b) − F(a): TOP minus bottom, 9 − 1.
  • A sign check catches it: 2x is positive on the whole interval, so the integral must be positive.
  • 9 − 1 = 8. It is also the area of a trapezoid with heights 2 and 6 and width 2.

Step through the notes. The antiderivative is right, and the arithmetic is right, but the subtraction is the wrong way round. Part 2 is the value at the top minus the value at the bottom.

The sign check is the fastest way to catch this. The integrand, 2x, is positive everywhere between 1 and 3, so every Riemann sum rectangle is above the axis, and the integral must be positive. A negative answer is impossible.

The region is a trapezoid with parallel sides 2 and 6 and width 2, so its area is 8 by geometry alone. Using a second method on easy integrals like this builds the instinct that catches sign errors on hard ones.

52. Trap: Part 2 across a discontinuity

Trap

The trap

A positive integrand, a negative answer:

\[ \int_{-1}^{1}\frac{dx}{x^2} \]

\[ = \left[-\frac1x\right]_{-1}^{1} = -2 \]

Wrong. The hypothesis fails.

The fix

The integrand blows up at 0, inside the interval, so it is not continuous there and Part 2 does not apply. The region has infinite area (an improper integral, Section 3.7 of Volume 2).

\[ \frac{1}{x^2} > 0 \text{ yet } -2 < 0 \]

A sign that cannot be right is the warning.

The warning sign is the answer itself: one over x squared is positive everywhere it is defined, so its integral over any interval cannot be negative. When Part 2 produces an impossible answer, check its hypothesis.

The integrand is not continuous on the interval from minus 1 to 1; it is not even defined at 0, and it grows without bound near 0. Part 2 needs continuity on the whole closed interval, so it simply does not apply. Before evaluating, always scan the interval for points where the denominator is zero.

53. What the theorem means

Section

Part 5

54. Integrating a rate gives the net change

Concept

Part 2 applied to a derivative: if F prime is a rate, its integral is how much F changed.

\[ \int_a^b F'(x)\,dx = F(b) - F(a) \]

rate F′its integral from a to b
velocity v(t)displacement s(b) − s(a)
flow rate (litres/min)change in volume
marginal cost C′(q)extra cost of raising output from a to b

Net means signed: time spent moving backward counts negative. The theorem says nothing about the total amount of motion.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 478 — the evaluation theorem in applications

Part 2 has a second reading that makes it useful far beyond areas. Any derivative is a rate of change. Integrating a rate over an interval gives the total change of the quantity over that interval.

Look at the table. Integrate a velocity and you get a change in position. Integrate a flow rate and you get a change in volume. Integrate marginal cost and you get the extra cost of producing more. It is one theorem in three costumes.

The word net is important. If the rate is sometimes negative, those stretches subtract. A car that drives forward 10 miles and back 4 has a net change in position of 6 miles, not 14. Section 5.4 develops this idea as the Net Change Theorem.

55. Example 5.22: the roller-skating race

Worked example

James skates at 5 plus 2t ft/s and Kathy at 10 plus cos of pi t over 2 ft/s. Who has gone further after 5 seconds?

\[ f(t) = 5 + 2t, \qquad g(t) = 10 + \cos\left(\tfrac{\pi}{2}t\right) \]

James: integrate his velocity

Why: Power rule.

\[ \int_0^5 (5+2t)\,dt = \big(5t + t^2\big)\Big|_0^5 = 25 + 25 = 50 \]

Kathy: an antiderivative of the cosine

Why: The chain rule would produce a factor pi over 2, so divide by it.

\[ \int \cos\left(\tfrac{\pi}{2}t\right)dt = \frac{2}{\pi}\sin\left(\tfrac{\pi}{2}t\right) \]

Kathy: evaluate

Why: The sine of 5 pi over 2 is 1, and the sine of 0 is 0.

\[ \left(10t + \tfrac{2}{\pi}\sin\tfrac{\pi t}{2}\right)\Big|_0^5 = 50 + \tfrac{2}{\pi} \approx 50.64 \]

Figure (svg): Velocity curves for the race from t equals 0 to 5: James's straight line 5 plus 2t and Kathy's wavy curve 10 plus cos of pi t over 2. Kathy is faster until t equals 2, where both skate at 9 ft/s (the gap shaded in her colour), James faster afterwards (shaded in his).

Distance is area under velocity. Kathy's early lead (orange) is slightly bigger than James's late comeback (blue), which is why she wins by about 0.64 ft.

Check by differentiating Kathy's antiderivative

Why: It must give back her velocity; then compare the totals: Kathy wins by 2 over pi, about 0.64 ft.

\[ \frac{d}{dt}\left(10t + \tfrac{2}{\pi}\sin\tfrac{\pi t}{2}\right) = 10 + \cos\tfrac{\pi t}{2} \;\checkmark, \quad 50.64 > 50 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 481 — Example 5.22

The question is who has gone further, so you integrate each velocity over the race. James's velocity is a line, and the power rule gives 50 feet.

Kathy's velocity has a cosine of pi t over 2. The derivative of the sine of pi t over 2 has an extra factor of pi over 2 from the chain rule, so the antiderivative must divide by it: that is where the 2 over pi comes from. The sine of five pi over 2 is 1, so she covers 50 plus 2 over pi, about 50.64 feet.

Now look at the figure. The distance each skater covers is the area under their velocity curve, so the difference between them is the area between the curves. Kathy is faster until exactly 2 seconds, when both are skating at 9 feet per second, and James after, and Kathy's early advantage, the orange area, is slightly bigger than James's comeback. She wins by about two thirds of a foot.

56. Checkpoint 5.20: the rematch stops at 3 seconds

Worked example

Same velocities, but the race ends at t equal to 3. Does the result change?

James

Why: His antiderivative at 3.

\[ \big(5t + t^2\big)\Big|_0^3 = 15 + 9 = 24 \]

Kathy

Why: The sine of 3 pi over 2 is minus 1.

\[ \left(10t + \tfrac{2}{\pi}\sin\tfrac{\pi t}{2}\right)\Big|_0^3 = 30 - \tfrac{2}{\pi} \approx 29.36 \]

Compare

Why: Kathy is further ahead than before.

\[ 29.36 > 24 \]

Check against the picture

Why: James only becomes faster at t equal to 2, when Kathy's lead peaks at 6 ft. One second later he has won back just 0.64 ft of it. The outcome is the same.

\[ \text{lead at } 2: 20 - 14 = 6, \qquad \text{lead at } 3: 29.36 - 24 = 5.36 \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 481 — Checkpoint 5.20

Work it the same way with 3 as the upper limit. James covers 15 plus 9, which is 24 feet. For Kathy, the sine of three pi over 2 is minus 1, so she covers 30 minus 2 over pi, about 29.36 feet.

Kathy still wins, and by much more: over five feet. It might seem odd that she wins by more in a shorter race. Look back at the race figure: James only becomes faster at 2 seconds, when Kathy's lead has reached 6 feet, so by 3 seconds he has won back less than a foot of it.

The general lesson is that a comparison of distances is a comparison of areas, and stopping the race earlier cuts off part of the area. Always think about which stretch of the graph each answer includes.

57. Net change or total distance? A velocity that changes sign

Worked example

A particle moves along a line with the velocity below, in metres per second. Find its displacement and the distance it travels in the first 3 seconds.

\[ v(t) = t^2 - 4, \qquad 0 \le t \le 3 \]

Displacement: integrate the velocity

Why: Part 2 gives the net change in position.

\[ \int_0^3 (t^2-4)\,dt = \left(\frac{t^3}{3} - 4t\right)\Big|_0^3 = 9 - 12 = -3 \]

Find where v changes sign

Why: Solve v equal to 0 inside the interval.

\[ t^2 - 4 = 0 \;\Longrightarrow\; t = 2 \]

Backward leg, from 0 to 2

Why: v is negative, so use its opposite.

\[ \int_0^2 (4-t^2)\,dt = 8 - \frac83 = \frac{16}{3} \]

Forward leg, from 2 to 3

Why: v is positive.

\[ \int_2^3 (t^2-4)\,dt = (9 - 12) - \left(\frac83 - 8\right) = \frac73 \]

Figure (svg): The velocity v of t equals t squared minus 4 from 0 to 3. From 0 to 2 the curve is below the axis and that region, area 16 over 3, is shaded red; from 2 to 3 it is above the axis and that region, area 7 over 3, is shaded green.

Backward for two seconds, forward for one. The net change adds the signed areas; the distance adds their sizes.

Check: the signed legs must give the displacement

Why: Backward sixteen thirds, forward seven thirds; the distance is their sum.

\[ -\frac{16}{3} + \frac73 = -3 \;\checkmark, \qquad \frac{16}{3} + \frac73 = \frac{23}{3} \approx 7.67 \text{ m} \]

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 478 — the evaluation theorem

This is the classic place where net change and total differ. Integrating the velocity from 0 to 3 gives minus 3 metres: after 3 seconds the particle is 3 metres behind where it started. That is the displacement.

But the particle did not simply move 3 metres. The velocity is negative until t equal to 2, so it first moves backward, then turns round and moves forward. To get the distance travelled, split at the turning point and add the sizes of the two legs.

The figure shows the two legs as areas: sixteen thirds backward in red, seven thirds forward in green. The check adds them with signs, which must give the displacement again, and without signs, which gives the distance, twenty-three thirds, about 7.67 metres.

58. Exercise 194: remove the absolute value

Fill the middle

Split at the root of the integrand, x equal to 0, and integrate each piece. Fill in the two areas and the total.

Fill in the blanks

\int_2^\frac{9}{2}|x|\,dx = \int_\frac{13}{2}^___(-x)\,dx + \int_0^3 x\,dx = ___ + ___ = ___

Why: On the left piece the absolute value of x is minus x, and minus x squared over 2 from minus 2 to 0 gives 2. On the right, x squared over 2 from 0 to 3 gives 9 over 2. Both are triangle areas, and together they make 13 over 2.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, p. 485 — Exercise 194

The absolute value of x has no single antiderivative formula that works on both sides of zero, so split the interval where the expression inside changes sign. On the left of zero, the absolute value of x is minus x; on the right, it is x.

Each piece is a simple power-rule integral, and each is also a triangle, which gives you a quick check: a triangle with base 2 and height 2 has area 2, and one with base 3 and height 3 has area nine halves.

The same strategy works for any absolute value: find the roots of the expression inside, split there, and remove the bars with the correct sign on each piece. It is exactly the distance calculation from the previous slide in a different setting.

59. The two parts side by side

Comparison

Comparison matrix

Part 1Part 2
the formulad/dx ∫ₐˣ f(t) dt = f(x)∫ₐᵇ f(x) dx = F(b) − F(a)
the limits of integrationa and the variable xtwo numbers, a and b
the result is a…functionnumber
needs an antiderivative in closed form?noyes
which operation comes firstintegrate, then differentiatedifferentiate (to find F), then integrate

Fill in the blanks by asking what each part takes in and gives out. Part 1 has a variable upper limit, so its integral is a function of x, and differentiating it needs no antiderivative in closed form: you only need the integrand.

Part 2 has two numbers as limits, so its integral is a number, and computing it does need an antiderivative you can write down. That is its limitation: when no formula exists, as for e to the minus x squared, Part 2 cannot help and you are back to numerical sums.

Look at the last row. The two parts undo the two operations in opposite orders. That pair of facts is what it means for differentiation and integration to be inverse processes.

60. A skydiver's free fall

Real world

The section's closing project. Julie falls with velocity 32t ft/s until she reaches her belly-down terminal velocity of 176 ft/s, then falls at that constant speed.

Discussion prompt

How long until she reaches terminal velocity, and how far has she fallen after 30 seconds? Set it up as integrals first.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 482-483 — Student Project: A Parachutist in Free Fall

This is the project at the end of the section. Set it up before calculating. Julie's velocity is 32t until it reaches 176, then it stays at 176. Solving 32t equal to 176 gives 5.5 seconds.

Because the velocity follows two different formulas, the distance integral splits at 5.5 seconds: the area under the rising line, a triangle, plus the area under the flat line, a rectangle. That gives 484 feet plus 4312 feet, 4796 feet in the first 30 seconds.

Notice that the answer is just triangle plus rectangle, which you could have found without calculus. The integrals matter because they say which areas to add, and because the same setup works unchanged when the velocity is curved, as it is for a real skydiver with air resistance.

61. Why it is called fundamental

Intuition

Before Newton and Leibniz, tangents and areas were separate subjects with separate tricks, and areas were by far the harder one. The theorem says the two operations undo each other.

\[ \frac{d}{dx}\int_a^x f(t)\,dt = f(x) \qquad\text{and}\qquad \int_a^b F'(x)\,dx = F(b) - F(a) \]

Part 1: integrate, then differentiate, and you are back to f. Part 2: differentiate, then integrate, and you are back to F, up to where you started. Every table of derivatives is secretly a table of integrals.

OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 472-473 and 476 — introduction and remarks after Theorem 5.4

For two thousand years, finding areas was hard. Archimedes found the area of a parabolic segment with an ingenious argument that worked for that one shape. Finding tangents, the derivative problem, was a separate subject. Newton and Leibniz, in the late 1600s, saw that the two problems are inverse to each other.

The two formulas on the slide say this exactly. Part 1: integrate f and then differentiate, and you get f back. Part 2: differentiate F and then integrate, and you get the change in F back.

The practical consequence is enormous. Every derivative you know, read backwards, is an integral you can now evaluate. That is why the next section is essentially a table of derivatives turned into a table of integrals.

62. Putting it together

Section

Part 6

63. Pattern: which part of the theorem?

Pattern

Figure (svg): A flow diagram. Start: is x in a limit of integration? Yes leads to Part 1: plug the limit into f and multiply by its derivative, subtracting for a lower limit. No, the limits are numbers, leads to: is f continuous on the interval? If not, stop, Part 2 does not apply. If so, find an antiderivative F and compute F of b minus F of a.

One question sorts every problem in this section: is the variable in a limit, or are the limits numbers?
  1. Is x in a limit of integration? Then you are differentiating an integral: Part 1.
  2. Plug the limit into the integrand and multiply by the limit's derivative. A lower limit gets a minus sign; two moving limits give top minus bottom.
  3. Are both limits numbers? Then you are evaluating: Part 2. First check that f is continuous on the whole interval.
  4. Rewrite into powers, find an antiderivative, drop the constant, compute top minus bottom.
  5. Check the sign and size against a sketch, and say whether a question wants a net change or a total.

Start every problem in this section with the question in the first box of the diagram: where is the variable? If x appears in a limit of integration, you are being asked for a derivative of an integral, and the answer comes from Part 1 with the chain rule. No integrating is needed.

If both limits are numbers, you are being asked for a number, and the tool is Part 2. The first thing to check is continuity on the whole interval, then rewrite the integrand into powers or recognisable derivatives before looking for an antiderivative.

The last step in the list is the one that separates a correct answer from a nearly correct one: check the sign against a sketch, and decide whether the question wants a net change or a total. The theorem computes net change, and totals need the interval split.

64. Put an evaluation in order

Ranking

Put in order

Order the steps for evaluating the integral from 1 to 9 of (x − 1)/√x.

  1. Confirm the integrand is continuous on [1, 9]
  2. Rewrite as powers: x^(1/2) − x^(−1/2)
  3. Find an antiderivative F, taking C = 0
  4. Compute F(9) and F(1)
  5. Subtract F(9) − F(1) and check the sign and size

Why: The hypothesis comes first, because Part 2 is meaningless without it. Rewriting comes before the power rule can apply, and the subtraction is top minus bottom.

Arrange the steps, then compare with Example 5.21, which is exactly this integral. The continuity check comes first: if the integrand had a problem inside the interval, nothing after it would be valid.

Rewriting into powers must come before the antiderivative, because the power rule only applies to powers. Evaluating at both ends and subtracting, top minus bottom, finishes the calculation.

The final sign-and-size check is part of the method, not an optional extra. Here the integrand is positive from 1 to 9, so the answer must be positive, and a region 8 wide and under 3 tall must have area under 24.

65. Check: a variable upper limit

Check

Check your understanding

What is d/dx of the integral from 2 to x³ of sin t dt?

  • A. 3x² sin(x³) (correct)
  • B. sin(x³)
  • C. sin(x³) − sin 8
  • D. −3x² cos(x³)

Answer: A

Why: Put the upper limit into the integrand, giving sin(x³), and multiply by the derivative of the limit, 3x². The constant lower limit contributes nothing.

Why B tempts people
This forgets the chain-rule factor: the limit is x³, not x, so you must multiply by 3x².
Why C tempts people
The lower limit is a constant; subtracting its value is Part 2 thinking, and the derivative of a constant is zero anyway.
Why D tempts people
This differentiates the integral as if it were already an antiderivative. Part 1 needs no integrating: plug the limit in.

Apply the recipe: the upper limit is x cubed, so put x cubed into the sine and multiply by the derivative of x cubed. The constant lower limit contributes nothing.

If you chose an answer without the 3x squared, look again at the chain-rule trap earlier in the lesson. If you chose one with a cosine, you integrated the sine; Part 1 never requires integrating. And if you subtracted something for the lower limit, remember that the lower limit only adds a constant, which differentiates to zero.

66. Check: evaluating with Part 2

Check

Check your understanding

What is the integral from 0 to π of sin x dx?

  • A. 0
  • B. 2 (correct)
  • C. −2
  • D. 1

Answer: B

Why: An antiderivative of sin x is −cos x. Top minus bottom: −cos π − (−cos 0) = 1 + 1 = 2. The sine is positive on the whole interval, so a positive answer is expected.

Why A tempts people
0 is the integral over a full period, 0 to 2π, where the positive and negative halves cancel. On 0 to π there is no negative half.
Why C tempts people
This uses cos x as the antiderivative, or subtracts bottom minus top. Either way the sign check fails: sin x is positive here.
Why D tempts people
1 is the height of the sine curve at its peak, not the area under the arch.

The antiderivative of sine is minus cosine, not cosine; getting that sign right is the whole question. Then top minus bottom: minus cosine of pi is 1, and minus cosine of 0 is minus 1, so the answer is 1 minus minus 1, which is 2.

The sine curve is above the axis on the whole interval from 0 to pi, so the answer has to be positive. That sign check rules out minus 2 immediately. And 0 is the answer over a full period, where the arch below the axis cancels the arch above.

67. Check: net change against distance

Check

Check your understanding

A particle has velocity v(t) = 2t − 4 m/s. How far does it travel (total distance) from t = 0 to t = 3?

  • A. −3 m
  • B. 3 m
  • C. 5 m (correct)
  • D. 9 m

Answer: C

Why: The velocity is negative until t = 2 and positive after. Backward leg: the integral of 4 − 2t from 0 to 2 is 4. Forward leg: the integral of 2t − 4 from 2 to 3 is 1. Distance 4 + 1 = 5 m, while the displacement is 1 − 4 = −3 m.

Why A tempts people
−3 is the displacement, the net change in position. Distance counts backward motion as positive.
Why B tempts people
This is the size of the displacement. Taking the absolute value at the end is not the same as adding the sizes of the two legs.
Why D tempts people
9 is t² at t = 3; the antiderivative of 2t − 4 also includes −4t.

The question asks for distance, so you cannot simply integrate the velocity. First find where it changes sign: 2t minus 4 is zero at t equal to 2.

Before 2 seconds the particle moves backward a distance of 4 metres, the area of a triangle with base 2 and height 4. After that it moves forward 1 metre. The distance is 4 plus 1, which is 5 metres, while the displacement, the plain integral, is minus 3. The distractors are exactly the ways of mixing those two ideas up.

68. Explain the inverse relationship

Explain it to yourself

Discussion prompt

In two or three sentences, explain why the rate at which the area under f grows, as the right edge moves, is exactly the height f(x). Use the thin strip, and say where continuity is used.

Write your explanation before revealing the model answer. A good answer uses the strip picture and says clearly where continuity comes in.

The key sentence is that the strip's area divided by its width is an average height, and the Mean Value Theorem for Integrals turns an average height into an actual height at some point. As the strip narrows, that point is forced onto x.

If your explanation said only that the strip is roughly a rectangle, that is the right intuition but not yet a proof. The word roughly is what the Mean Value Theorem for Integrals and continuity make exact.

69. Exit ticket

Exit ticket

Discussion prompt

Do both, and say which part of the theorem each one uses.

Do both parts before you look, and label each with the part of the theorem it uses. Part (a) has x squared as its upper limit, so it is a Part 1 question with the chain rule: put x squared into one over t and multiply by 2x, which simplifies to 2 over x.

Part (b) has numbers as both limits, so it is a Part 2 question. The antiderivative of 3t squared minus 2t is t cubed minus t squared; at 3 it is 18 and at 1 it is 0, so the integral is 18.

The check on part (a) is worth noticing: the integral of one over t from 1 to x squared is the natural logarithm of x squared, which is 2 ln x, and its derivative is 2 over x. If both answers came out right, you have the core skills of this section.

70. Recap

Recap

resultstatementuse it to
Mean Value Theorem for Integralsf(c)(b − a) = ∫ₐᵇ f dx for some cfind where f equals its average
Part 1d/dx ∫ₐˣ f(t) dt = f(x)differentiate an integral with a variable limit
with the chain ruled/dx ∫ₐ^u(x) f(t) dt = f(u(x)) u′(x)handle a function in a limit; bottom limits subtract
Part 2∫ₐᵇ f dx = F(b) − F(a)evaluate a definite integral from any antiderivative
net change∫ₐᵇ F′(x) dx = F(b) − F(a)turn a rate into a signed change

\[ \frac{d}{dx}\int_a^x f(t)\,dt = f(x), \qquad \int_a^b f(x)\,dx = F(b) - F(a) \]

Next, Section 5.4 builds the table of antiderivatives these evaluations need and applies the net change idea to motion, volume and cost.

Stewart, Calculus: Early Transcendentals 8e, §5.3 The Fundamental Theorem of Calculus §5.3, pp. 392-401 — the same material in Stewart

The table collects the five results of the lesson. The Mean Value Theorem for Integrals is the tool; Part 1 and its chain-rule extension differentiate integrals; Part 2 evaluates them; and the net change reading is Part 2 applied to rates.

If you remember only one line, remember the two formulas displayed under the table. Differentiation undoes integration, and integration undoes differentiation up to where you start. Every technique in the rest of this course is built on those two facts.

In the next section you will turn your table of derivatives into a table of integrals and use the net change idea on motion and other rates. Keep the sign check habit: it is the fastest way to catch the most common errors.

Sources

  1. OpenStax Calculus Volume 1, §5.3 The Fundamental Theorem of Calculus — Strang & Herman, OpenStax / Rice University, CC BY-NC-SA 4.0, pp. 472-487
  2. Stewart, Calculus: Early Transcendentals 8e, §5.3 The Fundamental Theorem of Calculus — James Stewart, Cengage Learning, 2016, pp. 392-401

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

Book on Wyzant · Text (657) 465-8108