1.4. Abstract measure spaces 89 Let 0 ε 1 be arbitrary. Then we have f(x) = sup n fn(x) (1 − ε)ci for all x ∈ Ai. Thus, if we define the sets Ai,n := {x ∈ Ai : fn(x) (1 − ε)ci} then the Ai,n increase to Ai and are measurable. By upwards monotonicity of measure, we conclude that lim n→∞ μ(Ai,n) = μ(Ai). On the other hand, observe the pointwise bound fn ≥ k i=1 (1 − ε)ci1Ai,n for any n integrating this, we obtain X fn dμ ≥ (1 − ε) k i=1 ciμ(Ai,n). Taking limits as n → ∞, we obtain lim n→∞ X fn dμ ≥ (1 − ε) k i=1 ciμ(Ai) sending ε → 0 we then obtain the claim. Remark 1.4.44. It is easy to see that the result still holds if the mono- tonicity fn ≤ fn+1 only holds almost everywhere rather than everywhere. This has a number of important corollaries. First, we can generalise (part of) Tonelli’s theorem for exchanging sums (see Theorem 0.0.2): Corollary 1.4.45 (Tonelli’s theorem for sums and integrals). Let (X, B,μ) be a measure space, and let f1,f2,... : X → [0, +∞] be a sequence of un- signed measurable functions. Then one has X ∞ n=1 fn dμ = ∞ n=1 X fn dμ. Proof. Apply the monotone convergence theorem (Theorem 1.4.43) to the partial sums FN := ∑N n=1 fn. Exercise 1.4.42. Give an example to show that this corollary can fail if the fn are assumed to be absolutely integrable rather than unsigned measurable, even if the sum ∑∞ n=1 fn(x) is absolutely convergent for each x. (Hint: Think about the three escapes to infinity.)

Purchased from American Mathematical Society for the exclusive use of nofirst nolast (email unknown) Copyright 2011 American Mathematical Society. Duplication prohibited. Please report unauthorized use to cust-serv@ams.org. Thank You! Your purchase supports the AMS' mission, programs, and services for the mathematical community.