— Assume that G is locally compact. Let μ be a nonzero relatively invariant measure on X. Then its multiplier χ is a continuous function on G.
For, let f∈K(X), S the support of f, s0 a point of G, and V a compact neighborhood of s0 in G; then, the set
s∈V⋃γ(s)−1(S)=V−1S
is compact in X; by Lemma 1 and formula (8), χ(s−1)⟨μ,f⟩ depends continuously on s; if f is chosen so that ⟨μ,f⟩=0, one sees that χ is continuous.
Now let G be a topological group operating continuously on the right in a locally compact space X; for s∈G and x∈X, let xs be the transform of x by s. We denote by δX(s), or δ(s), the homeomorphism of X defined by
(1')
δ(s)x=xs−1.
We have
(2')
δ(st)=δ(s)δ(t).
By transport of structure, one defines the action of δ(s) on functions and measures on X:
(3')
(δ(s)f)(x)=f(xs) (4')
⟨f,δ(s)μ⟩=⟨δ(s−1)f,μ⟩ (5')
∫Xf(x)d(δ(s)μ)(x)=∫Xf(xs−1)dμ(x) (6')
(δ(s)μ)(A)=μ(As).
We agree to write dμ(xs) in place of d(δ(s)μ)(x), and (5') then takes the form
∫Xf(x)dμ(xs)=∫Xf(xs−1)dμ(x).
One defines in an analogous manner the measures on X invariant, relatively invariant and quasi-invariant under G. If μ is relatively invariant, its multiplier χ is defined by the formulas
(7′)δ(s)μ=χ(s)μ
(8′)∫Xf(xs)dμ(x)=χ(s)−1∫Xf(x)dμ(x)
(9′)μ(As)=χ(s)μ(A)
(10′)dμ(xs)=χ(s)dμ(x).
If one regards the group G0 opposite to G as operating in X by (x,s)↦xs, then μ is relatively invariant under G0 with the same multiplier χ.
Finally, let G be a locally compact group. It operates on itself by left and right translations, according to the formulas γ(s)x=sx, δ(s)x=xs−1. Then
(11)γ(s)δ(t)=δ(t)γ(s).
All of the foregoing is applicable here, thus we have on G the concepts of measures that are left-invariant, right-invariant, relatively left-invariant, relatively right-invariant, left quasi-invariant, right quasi-invariant (however, see Nos. 8 and 9).
The mapping x↦x−1 is a homeomorphism of G onto G. For every function f on G, we define the function fˇ on G by
(12)fˇ(x)=f(x−1).
For every measure μ on G, we define the measure μˇ by
(13)μˇ(f)=μ(fˇ)for f∈K(G).
In other words,
(14)∫Gf(x)dμˇ(x)=∫Gf(x−1)dμ(x).
If A is a μˇ-integrable set, then A−1 is μ-integrable and
(15)μˇ(A)=μ(A−1).
We agree to write dμ(x−1) in place of dμˇ(x), and (14) then takes the form
∫Gf(x)dμ(x−1)=∫Gf(x−1)dμ(x).
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.