The Fundamental Theorem of Calculus links derivatives and integrals. Part 1: if f is continuous and F(x)=∫axf(t) dt, then F′(x)=f(x). Part 2: if F is any antiderivative of a continuous f on [a,b], then the integral of f from a to b is F(b)−F(a).
A concrete case
Distance two ways
A car pulls away from a stop sign and speeds up steadily. Its speed starts at 0 m/s and rises by 2 m/s every second, so after 3 seconds it is doing 6 m/s. How far has it gone?
One way reads the answer from the graph of speed against time. At a constant speed, distance is speed times time, which is the area of a rectangle under the speed graph. For a changing speed the distance is still the area under the graph. Here that area is a triangle 3 s wide and 6 m/s tall, so the car covers 21⋅3⋅6=9 m.
The other way uses a distance formula. The speed after t seconds is 2t, and t2 has derivative 2t, so the distance is t2 meters. That gives 1,4 and 9 m at 1,2 and 3 seconds. The table shows the two routes agreeing second by second.
Now ask for the distance between t=1 and t=3. The area under the speed graph there is a trapezoid, (2+6)/2⋅2=8 m. The formula gives 32−12=8 m. The area under a rate graph equals the change in the quantity the rate measures. That is the Fundamental Theorem of Calculus.
22+6⋅2=8=32−12
The car’s distance, second by second
Second
Speed at its start and end (m/s)
Area under the speed graph (m)
Change in t2(m)
0 to 1
0 and 2
(0+2)/2⋅1=1
1−0=1
1 to 2
2 and 4
(2+4)/2⋅1=3
4−1=3
2 to 3
4 and 6
(4+6)/2⋅1=5
9−4=5
The rule
Part 1: the area function grows at rate f
Fix a starting point a, and let f be continuous. For each x, let A(x) be the signed area under the graph of f from a to x, written ∫axf(t) dt.A is called an accumulation function. The letter t is a placeholder for the variable of integration. The variable that matters is x, the moving right edge.
Part 1 of the Fundamental Theorem says that A has a derivative, and that its derivative is f itself. The area grows exactly as fast as the curve is tall. Where f is negative, new area counts as negative and A falls.
A(x)=∫axf(t)dt⟹A′(x)=f(x)
A(x)=∫axf(t)dtA′(x)=f(x)
Part 2: evaluate with any antiderivative
Part 2 turns the theorem into a way of computing. Suppose f is continuous on [a,b] and F is any antiderivative of f there, meaning F′(x)=f(x). Then the integral of f from a to b is F(b)−F(a). Evaluate the antiderivative at the upper limit and subtract its value at the lower limit.
The difference is often written as F(x) in square brackets, with a and b at the foot and top of the closing bracket. Any antiderivative gives the same answer. A constant C added to F appears in both terms and cancels.
Books do not agree on the numbering. Some swap Part 1 and Part 2, and some call them the First and Second Fundamental Theorems. One statement differentiates an area function, and the other evaluates an integral with an antiderivative.
∫abf(x)dx=[F(x)]ab=F(b)−F(a)
∫abf(x)dx=[F(x)]ab=F(b)−F(a)
The integral of a rate is a net change
Part 2 has a reading that needs no area at all. If Q(t) is any quantity and Q′(t) is its rate of change, integrating the rate gives the change in the quantity. Integrating a velocity gives a displacement, and integrating a flow rate gives the net amount added. The units multiply, so liters per minute times minutes gives liters.
∫abQ′(t)dt=Q(b)−Q(a)
Why it works
Why the area grows at rate f
Take f(t)=t2 and A(x)=∫0xt2dt. As x moves from 1 to 1.01,A gains a thin strip 0.01 wide. Across that strip the curve is at least f(1)=1 tall and at most f(1.01)=1.0201 tall. The strip’s area therefore lies between 0.01 and 0.010201, and its area divided by its width lies between 1 and 1.0201. The table narrows the strip further.
The same squeeze works at any x and for any small width h. The gain A(x+h)−A(x), divided by h, is the average height of f across the strip. It lies between the lowest and highest values of f there. As h shrinks to 0, a continuous f pins both of those values to f(x), and the average height becomes A′(x)=f(x).
Part 2 follows from Part 1. A and any antiderivative F have the same derivative, f. By the Mean Value Theorem, two functions with the same derivative on an interval differ by a constant, so A(x)=F(x)+C. No area has built up at x=a, so 0=F(a)+C and C=−F(a). At x=b this reads A(b)=F(b)−F(a).
hA(x+h)−A(x)=h1∫xx+hf(t)dth→0f(x)
hA(x+h)−A(x)=h1∫xx+hf(t)dth→0f(x)
Squeezing the strip under t2 at x=1
Strip width h
Lowest height, f(1)
Highest height, f(1+h)
Strip area÷h
0.1
1
1.21
1.10333
0.01
1
1.0201
1.01003
0.001
1
1.002001
1.00100
The method
Evaluating a definite integral, step by step
The table evaluates ∫12 (6x2−4x) dx. Leave out the +C, which would cancel in the subtraction anyway.
∫12(6x2−4x)dx=[2x3−2x2]12=(16−8)−(2−2)=8
∫12(6x2−4x)dx=[2x3−2x2]12=(16−8)−(2−2)=8
Part 2 on the integral of 6x2−4x from 1 to 2
Step
What to do
For ∫12 (6x2−4x) dx
1
Check that f is continuous on [a,b]
6x2−4x is a polynomial, continuous everywhere
2
Find an antiderivative F
F(x)=2x3−2x2
3
Check F by differentiating
F′(x)=6x2−4x
4
Evaluate F at both limits
F(2)=16−8=8 and F(1)=2−2=0
5
Subtract, upper limit first
8−0=8
6
Read the sign and the units
Positive, since 6x2−4x>0 on [1,2]
Differentiating an accumulation function
Part 1 differentiates an accumulation function without integrating anything. Put the upper limit in place of t. When the upper limit is a function g(x) rather than x, the chain rule also multiplies by g′(x). When x is the lower limit, swap the limits first, which changes the sign of the integral.
For G(x)=∫0x2costdt, the integrand at the upper limit is cos(x2), and the upper limit has derivative 2x.G′(x)=2xcos(x2). Integrating first gives the same answer. G(x)=sin(x2)−sin0=sin(x2), and the chain rule gives 2xcos(x2) again.
dxd∫0x2costdt=cos(x2)⋅2x
Derivatives of accumulation functions
Function of x
Its derivative
∫ from a to x of f(t) dt
f(x)
∫ from a to g(x) of f(t) dt
f(g(x))⋅g′(x)
∫ from x to b of f(t) dt
−f(x)
Average value
Dividing an integral by the length of its interval gives the average value of the function. It is the height of the rectangle on [a,b] with the same area as the region under the curve. The Mean Value Theorem for integrals says that a continuous f takes its average value at least once in the interval.
favg=b−a1∫abf(x)dx
Where it breaks
Where it breaks
Part 2 needs f to be continuous on the whole of [a,b], and F to be an antiderivative on the whole of it. Take the integral of 1/x2from −1to 1. The function −1/x has derivative 1/x2, and subtracting blindly gives (−1)−1=−2. A positive function cannot have a negative integral. The trouble is at x=0, where 1/x2 is undefined and grows without bound. This integral is improper, and it diverges.
Part 1 needs f to be continuous at the point. Let f(t) be 0 for t<1 and 1 for t≥1, and A(x)=∫0xf(t) dt.A is 0 up to x=1 and x−1 after it. Its graph has a corner at 1, where A has no derivative. Everywhere else A′(x)=f(x).
The integral of a velocity is a displacement, not the distance traveled. Where the velocity is negative the object moves backward, and that stretch subtracts. For the distance, integrate the speed ∣v(t)∣ instead.
A definite integral and an indefinite integral are different objects. ∫122xdx is the number 3, and ∫2xdx is the family x2+C. The theorem connects them without making them the same thing.
∫−11x21dx=[−x1]−11=−2
Check your answer
Check your answer
An answer from the Fundamental Theorem can be checked backward, by its size and at its starting point.
Differentiate your antiderivative. F′(x) should give back the integrand exactly. For 2x3−2x2 it gives 6x2−4x.
Estimate the size. The integral lies between the smallest value of f times (b−a) and the largest value times (b−a). On [1,2],6x2−4x rises from 2 to 16, so ∫12 (6x2−4x) dx lies between 2and 16. One midpoint rectangle, f(1.5)⋅1=7.5, lands close to the exact 8.
Check the starting value. An accumulation function is 0 at its lower limit, because no area has built up yet. F(x)=x2/2−x passes, since F(0)=0.
To have Chalk check the work, write it in a Math box. A definite integral can share one row with its evaluation, as in ∫12 (6x2−4x) dx=(16−8)−(2−2)=8, and the Checker marks the row ✓ or ✗. The square bracket can share that row too, as in ∫12 (6x2−4x) dx=[2x3−2x2]12=8, which reads ✓. A wrong bracket reads ✗, and so does a right bracket with a wrong value. The bracket also reads on a row of its own, below the integral.
For an accumulation function, write its definition F(x)=∫… on the first row and the formula you found under it. The Checker marks that formula ✓ or ✗, and the derivative and the values you write under it are marked against it. Part 1 is checked directly too. Under G(x)=∫0x2costdt, the derivative G′(x)=2xcos(x2) reads ✓, and G′(x)=cos(x2), without the chain rule factor, reads ✗.
A proof from sums, and functions with no formula
Part 2 can be proved straight from Riemann sums. Split [a,b] at the points x0,x1, …, xn. The change F(b)−F(a) is the sum of the small changes F(xi)−F(xi−1), because each middle value appears once with each sign and cancels. By the Mean Value Theorem, each small change equals f(ci)Δx for some ci in its piece. F(b)−F(a) is therefore a Riemann sum of f, for every n. A number equal to every one of the sums equals their limit, the integral.
Part 1 says something Part 2 cannot. Every function that is continuous on an interval has an antiderivative there, even when no formula for one exists. The function e−t2 has no antiderivative built from the usual functions, but A(x)=∫0xe−t2dt is one all the same. Scaled by 2/√π, it is the error function that statistics uses for the normal curve.
Some books define the natural logarithm this way, as lnx=∫1x (1/t) dt for x>0. Part 1 then gives its derivative, 1/x, at once.
Subtracting in the wrong order. The integral from a to b is F(b)−F(a), upper limit first. ∫12 (6x2−4x) dx is 8−0=8, and 0−8 has the right size and the wrong sign.
Forgetting the chain rule factor. The derivative of ∫0x2costdt is 2xcos(x2), not cos(x2), because the upper limit x2 moves 2x times as fast as x.
Using Part 2 across a point where f blows up. The integral of 1/x2from −1 to 1 is not −2. The integrand is unbounded at 0, so the theorem does not apply, and the integral diverges.
Confusing the area function with its slope. For F(x)=∫0x (t−1) dt,F(2)=0 is the area built up so far, and F′(2)=1 is how fast it is growing. A zero area can still be growing.
Reading a net change as a total. A velocity whose integral is 0 can still cover ground. v(t)=t2−4t+3 has displacement 0 on [0,3], yet the particle travels 8/3 m, out and back.
Leaving the variable of integration in the answer. The derivative of ∫1xt2dt is x2, not t2. The t is a placeholder, and it disappears once the limit x is put in its place.
Key terms
Fundamental theorem of calculus
The link between derivatives and integrals: if F is an antiderivative of a continuous f, then ∫abf(x) dx=F(b)−F(a). It also says d/dx∫axf(t) dt=f(x).
Accumulation function
A function defined by an integral with a variable upper limit, F(x)=∫axf(t) dt. It records the signed area from a to x, and where f is continuous, F′(x)=f(x).
Net change theorem
The integral of a rate of change over an interval equals the net change in the quantity: ∫abF′(t) dt=F(b)−F(a). Decreases count as negative, so net change can differ from the total amount of change.
Antiderivative
A function whose derivative is the given function, such as x3 for 3x2. Any two antiderivatives on one interval differ by a constant, which is why answers carry +C.
Definite integral
∫abf(x) dx: the signed area between the graph of f and the x-axis from a to b, defined as the limit of Riemann sums. Area below the axis counts as negative.
Average value of a function
The integral of f over [a,b] divided by b−a: the height of the rectangle on [a,b] with the same signed area as f. A continuous function reaches its average value somewhere on the interval.
Work through an example
Let F(x)=∫0x (t−1) dt. Find a formula for F and its derivative, and explain where F decreases and why F(2)=0.