If t \in \mathscr{T}^{(\infty)}(G) and g \in G, then γ(g)∗t=εg∗t,δ(g)∗t=t∗εg−1, (Int g)_* t = εg∗t∗εg−1. In particular, εe is the unit element of \mathscr{T}^{(\infty)}(G).
Consider the diagram
GϕG×GψG
where ϕ is the mapping h↦(g,h) and ψ is the mapping (h′,h)↦h′h. Then γ(g)=ψ∘ϕ and hence γ(g)∗t=ψ∗(ϕ∗(t)). But ϕ∗(t)=εg⊗t and hence ψ∗(ϕ∗(t))=εg∗t. The argument is similar for δ(g)∗t. Finally,
Intg=γ(g)∘δ(g)
and hence (Intg)∗=γ(g)∗∘δ(g)∗.
It is therefore seen that, for t∈T(G), εg∗t and t∗εg are equal to gt and tg calculated in the group T(G) (§ 2, no. 2). But it should be noted that, for t,t′ in T(G), the product tt′ in the sense of § 2 is in general different from t∗t′.
Bourbaki puts the proof directly after the statement, so what is above is the statement and its proof together.