— *Let T and B be two locally compact spaces having countable bases, μ a positive measure on T, p a μ-measurable mapping of T into B, and ν a pseudo-image measure of μ under p. Then there exists a ν-adequate family b↦λb (b∈B) of positive measures on T, having the following properties: a) λb=0 for b∈p(T); b) λb is concentrated on the set p−1(b) for every b∈B; in particular, λb=0 for b∈/p(T); c) μ=∫λbdν(b).
Moreover, if ν′=r⋅ν is a second pseudo-image measure of μ under p, and if b↦λb′ is a ν′-adequate family of positive measures on T having the properties b) and c) with respect to ν′, then λb=r(b)λb′ almost everywhere in B (for ν or ν′).
There exists a continuous and finite numerical function f defined on T, such that f(t)>0 for every t∈T and such that μ′′=f⋅μ is bounded (Ch. V, §5, No. 6, Prop. 11). Let ν′′=p(μ′′), which is equivalent to ν, and write ν′′=g⋅ν, with g finite and locally ν-integrable; one can suppose, moreover, that g(b)>0 for all b∈B (Ch. V, §5, No. 6, Prop. 10). Th. 1 of No. 1, applied to μ′′ and ν′′, shows that there exists a ν′′-adequate family b↦λb′′ (b∈B) of positive measures on T, such that:
1) ∥λb′′∥=1 for b∈p(T); 2) λb′′ is concentrated on p−1(b) for every b∈B; 3) μ′′=∫λb′′dν′′(b).
For every b∈B, let us define a positive measure λb on T by the formula λb=(1/f)⋅(g(b)λb′′). It is clear that the family b↦λb has the properties a) and b) of the statement. On the other hand, for every function h∈K(T), h/f belongs to K(T), therefore
∫h(t)dμ(t)=∫(h(t)/f(t))dμ′′(t)=∫dν′′(b)∫(h(t)/f(t))dλb′′(t).
But since the function b↦∫(h(t)/f(t))dλb′′(t) is ν′′-integrable, the function b↦g(b)∫(h(t)/f(t))dλb′′(t) is ν-integrable (Ch. V, §5, No. 3, Th. 1).
By the definition of λb, this function is b↦∫h(t)dλb(t), whence (loc. cit.) ∫h(t)dμ(t)=∫dν(b)∫h(t)dλb(t), which proves that μ=∫λbdν(b).
To establish the second part of the statement, we remark that one can suppose that r(b)>0 for all b∈B (Ch. V, §5, No. 6, Prop. 10); set λb′′′=f⋅((r(b)/g(b))λb′); one shows, as above, that for every function h∈K(T), the relation
∫h(t)dμ(t)=∫dν′(b)∫h(t)dλb′(t)
implies
∫h(t)dμ(t)=∫dν′′(b)∫(h(t)/f(t))dλb′′′(t).
Therefore Th. 1 of No. 1, applied to μ′′ and ν′′, implies that for almost every b∈B, λb′′′=λb′′, whence λb=r(b)λb′.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.