Explanation
D/dx ∫ from g(x) to h(x) of f(t) dt = f(h(x))…
To differentiate an integral whose limits move, put the upper limit into the integrand and multiply by its derivative, then subtract the same thing for the lower limit.
d/dx ∫ from g(x) to h(x) of f(t) dt = f(h(x)) h′(x) − f(g(x)) g′(x)
Lim (1/n) Σ f(r/n) = ∫ from 0 to 1 of f(x) dx · r/n
A limit of a sum is an area. Write the sum as one over n times f of r over n.
Then r over n becomes x, one over n becomes dx, and the sum becomes the integral from zero to one.
lim (1/n) Σ f(r/n) = ∫ from 0 to 1 of f(x) dx · r/n → x, 1/n → dx
Lim Σ n/(n² + r²) = (1/n) Σ 1/(1 + (r/n)²)
Here's the trick in action. Divide top and bottom by n squared to get one over n times one over one plus r over n squared.
That's the integral of one over one plus x squared from zero to one: pi over four.
lim Σ n/(n² + r²) = (1/n) Σ 1/(1 + (r/n)²) → ∫₀¹ dx/(1 + x²) = π/4
Wallis
Wallis' formula handles powers of sin and cos from zero to pi by two. Multiply n minus one over n, n minus three over n minus two, and so on, down to one.
If n is even, finish with pi over two.
Wallis: ∫ from 0 to π/2 of sinⁿx dx = (n − 1)/n × (n − 3)/(n − 2) × … × (π/2 if n is even)
If m ≤ f(x) ≤ M on [a, b]
When you can't evaluate an integral, you can still trap it. If f stays between m and M, the area lies between m times the width and M times the width.
If m ≤ f(x) ≤ M on [a, b]: m(b − a) ≤ ∫ ≤ M(b − a)
Worked example
lim (1/n)[(1/n)² + (2/n)² + … + (n/n)²]. This is (1/n) Σ (r/n)², so it equals the integral from 0 to 1 of x² dx = 1/3.