Let (Eα,fαβ), (Fα,gαβ), (Gα,hαβ) be three inverse systems of sets relative to the same index set I; let E=←limEα, F=←limFα, G=←limGα, and let fα (resp. gα,hα) be the canonical mapping of E (resp. F,G) into Eα (resp. Fα,Gα). If (uα) and (vα) are two inverse systems of mappings, uα:Eα→Fα, vα:Fα→Gα, then the composite mappings vα∘uα:Eα→Gα form an inverse system of mappings, and we have
(7)
←lim(vα∘uα)=(←limvα)∘(←limuα).
For if we put wα=vα∘uα, then if α≤β we have
wα∘fαβ=vα∘(uα∘fαβ)=vα∘(gαβ∘vβ)=(hαβ∘vβ)∘uβ=hαβ∘wβ,
which shows that (wα) is an inverse system of mappings. Furthermore, if u=←limuα and v=←limvα, then hα∘(v∘u)=(vα∘gα)∘u=(vα∘uα)∘fα for each α∈I, and therefore, by the uniqueness of the inverse limit, we have v∘u=←limwα.
¶ Let (Eα,fαβ) be an inverse system of sets, and for each α∈I let Mα be a subset of Eα. If fαβ(Mβ)⊂Mα whenever α≤β, the Mα are said to form an inverse system of subsets of the Eα. Let gαβ be the mapping of Mβ into Mα (where α≤β) whose graph is the same as that of the restriction of fαβ to Mβ. Then it is clear that (Mα,gαβ) is an inverse system of sets and that
(8)
←limMα=(←limEα)∩α∈I∏Mα.
Bourbaki sets the proof directly after the statement, and the corpus does not separate them, so what is above is the statement and its proof together.
Nothing.