If the families (xλ)λ∈L and (yμ)μ∈M of finite real numbers are summable in R, then so is the family (xλyμ)(λ,μ)∈L×M and we have
(1)
(λ,μ)∈L×M∑xλyμ=(λ∈L∑xλ)(μ∈M∑yμ).
Every finite subset of L×M is contained in a finite subset of the form H×K, where H is a finite subset of L and K is a finite subset of M. By hypothesis, there exists a number a>0 such that ∑λ∈H∣xλ∣≤a and ∑μ∈K∣yμ∣≤a for all finite subsets H and K of L and M respectively; therefore
(λ,μ)∈H×K∑∣xλyμ∣=(λ∈H∑∣xλ∣)(μ∈K∑∣yμ∣)≤a2, and this shows that the family
(xλyμ) is summable in
R, by Theorems 1 and 3. By associativity we can write [Chapter III, § 5, no. 3, formula (2)]
(λ,μ)∈L×M∑xλyμ=λ∈L∑(μ∈M∑xλyμ)=λ∈L∑xλ(μ∈M∑yμ)=(λ∈L∑xλ)(μ∈M∑yμ); hence the result.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.