With notations as in Lemma 1, for every functionf∈K(X×Y,K×L;C)the function
y↦h(y)=∫f(x,y)dλ(x)
belongs toK(Y,L;C).
For every function u∈K(X;K(Y,L;C)), the integral ∫u(x)dλ(x) belongs to K(Y,L;C) since the latter is a Banach space (§ 3, No. 3, Cor. 1 of Prop. 7); but for u=ω(f) and for every y∈Y,
⟨∫u(x)dλ(x),εy⟩=∫u(x)(y)dλ(x)=∫f(x,y)dλ(x),
whence the lemma.
Consider, then, for every function f∈K(X×Y;C), the number ν(f)=μ(∫f(x,y)dλ(x)) (which we shall also write ∫dμ(y)∫f(x,y)dλ(x) by an abuse of notation); if K (resp. L) is a compact subset of X (resp. Y), then there exists a number aK (resp. bL) such that, for every function g∈K(X,K;C) (resp. h∈K(Y,L;C)), we have ∣λ(g)∣≤aK∥g∥ (resp. ∣μ(h)∣≤bL∥h∥). It follows that for every function f∈K(X×Y,K×L;C),
∫f(x,y)dλ(x)≤aK∥f∥
for every y∈Y, whence ∣ν(f)∣≤aKbL∥f∥. The linear form ν on K(X×Y;C) is thus a measure on X×Y that obviously satisfies (1), which completes the proof of Th. 1.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.