Upper and Lower Sums Close In

Fill rectangles under f(x)=x2f(x) = x^2. The lower and upper sums squeeze 13\frac{1}{3} from both sides.

L(f,P)≤∫abf≤U(f,P)L(f,P) \le \int_a^b f \le U(f,P)

Part of these learning paths

Examples & Exercises

Use ∑k=1nk2=n(n+1)(2n+1)6\sum_{k=1}^{n} k^2 = \frac{n(n+1)(2n+1)}{6} to prove ∫01x2 dx=13\int_0^1 x^2\,dx = \frac13 from upper and lower sums.

Ln=(n−1)n(2n−1)6n3→13,Un=n(n+1)(2n+1)6n3→13L_n = \frac{(n-1)n(2n-1)}{6n^3} \to \frac13, \quad U_n = \frac{n(n+1)(2n+1)}{6n^3} \to \frac13
n=48n = 48: the gap U−L=1/48≈0.0208U - L = 1/48 \approx 0.0208 squeezes 13\frac13.
Show proof
  1. On the uniform partition PnP_n, x2x^2 is increasing, so the lower sum uses left endpoints and the upper sum right endpoints: Ln=1n3∑k=0n−1k2L_n = \frac{1}{n^3}\sum_{k=0}^{n-1}k^2, Un=1n3∑k=1nk2U_n = \frac{1}{n^3}\sum_{k=1}^{n}k^2.
  2. Substituting the sum-of-squares formula gives Ln=(n−1)n(2n−1)6n3L_n = \frac{(n-1)n(2n-1)}{6n^3} and Un=n(n+1)(2n+1)6n3U_n = \frac{n(n+1)(2n+1)}{6n^3}.
  3. Both tend to 2n36n3=13\frac{2n^3}{6n^3} = \frac13 as n→∞n \to \infty, so the integral is 13\frac13 by squeezing.

Prove the Dirichlet function is not Riemann integrable by showing U−L=1U - L = 1 for every partition.

U(D,P)=1,L(D,P)=0for every PU(D,P) = 1, \quad L(D,P) = 0 \quad \text{for every } P
Contrast with the sums for x2x^2 — for the Dirichlet function the gap never narrows.
Show proof
  1. Every interval contains both rationals and irrationals, so on every subinterval sup⁡D=1\sup D = 1 and inf⁡D=0\inf D = 0.
  2. Hence for every partition PP, U(D,P)=∑1⋅Δxk=1U(D,P) = \sum 1\cdot\Delta x_k = 1 and L(D,P)=∑0⋅Δxk=0L(D,P) = \sum 0\cdot\Delta x_k = 0.
  3. Since U−L=1↛0U - L = 1 \not\to 0, the necessary and sufficient condition for Riemann integrability fails.

Prove that refining a partition never increases U−LU - L.

P′⊃P  ⇒  U(f,P′)≤U(f,P),  L(f,P′)≥L(f,P)P' \supset P \;\Rightarrow\; U(f,P') \le U(f,P),\; L(f,P') \ge L(f,P)
n=10n = 10: the gap shrinks monotonically as the partition refines.
Show proof
  1. Splitting one subinterval in two: each piece's supremum is at most the original's, so the upper sum decreases (or stays).
  2. Likewise each piece's infimum is at least the original's, so the lower sum increases (or stays).
  3. So U−LU - L decreases monotonically under refinement — the sums squeeze toward each other.

10. Integration